Abstract

Urban areas continue to be affected by flooding, necessitating more sustainable and effective adaptation strategies and mitigation initiatives. This study investigates the potential flood reduction capability achieved through implementing various green infrastructures known as low-impact development (LID). The Bendung watershed, in the center of Palembang City, with a total area of 18.37 km², is used as the study area to evaluate the performance of LID infrastructure in reducing flood parameters, including total runoff volume, peak runoff discharge, runoff coefficient, and flooding area. Five types of LID infrastructure were simulated, namely bio-retention cells, rain gardens, permeable pavements, rain barrels, and recharge wells. The flood simulations were performed using four design storms with 2-, 5-, 10-, and 25-year return periods. Hydrologic and hydraulic modeling and simulations were carried out using PCSWMM Professional 2D, and the results were integrated with ArcMap to map the flood inundation. The results of this study demonstrate that with only 9.81 percent of the area occupied by LIDs, a flood reduction of more than 30% can be achieved. In addition, implementing LIDs can help restore the watershed’s hydrological condition to its natural state, as indicated by the decrease in the runoff coefficient. Thus, implementing LIDs in a sustainable urban drainage system must be widely promoted in many urban areas, especially in developed countries like Indonesia. This study can be used as a reference for the local government and authorities to create policies and regulations to establish sustainable flood mitigation measures in Palembang City.

1 Introduction

The phenomenon of flooding is one of the main problems in urban areas. Rapid urbanization results in a wider impermeable area and reduced soil infiltration, with a consequent increase in runoff volume and a higher risk of flooding (Li et al. 2020). Currently, the pressures and risks on water systems in urban areas are increasing due to significant changes in the natural environment (Chen et al. 2016). Low-impact development (LID) techniques become an alternative strategy for controlling runoff and the risk of flooding by restoring the hydrological conditions of the built area to its natural state (Montazerolhodjah 2019; Zhao and Meng 2020). LIDs mimic the natural conditions of the catchment area designed to impede the flow of surface runoff, and even minimize it before it enters drainage channels, rivers, and other water bodies. The LID components provide storage for surface water, and more opportunities for rainwater to infiltrate the soil and be absorbed by plants before it evaporates from surface water and plants, known as evapotranspiration (Liu et al. 2021). Implementing LIDs is expected to increase a city’s resilience to floods and reduce the level of flood risk in urban areas (Arahuetes and Cantos 2019; Palermo et al. 2020).

Implementation of LIDs has been adopted in several countries and is known by different terminologies (Fletcher et al. 2015). The term Low-Impact Development (LID) is known and used in Canada and the United States (Nowogoński 2020). The term Water Sensitive Urban Design (WSUD) is also known and used in Australia. In addition, the term Sustainable Urban Drainage System (SUDS) is used in the United Kingdom (Fletcher et al. 2015). In several other studies, LIDs are also known as Green Infrastructure (GI) or Blue-Green Infrastructure (BGI) (Herath et al. 2023). Recently, China also adopted a similar concept for urban flood control known as Sponge City (Chan et al. 2018; Yin et al. 2022). Singapore uses the term Active, Beautiful, Clean (ABC) Waters, which refers to similar LID practices (Goh et al. 2017; Tan et al. 2019). LID components or infrastructures are divided into groups based on their infiltration and storage functions. The infiltration-based LID components include vegetative swales, infiltration trenches, recharge ponds, bio-retention cells or rain gardens, sand filter surfaces, and permeable pavement systems. The storage-based LID components are wetlands, retention ponds, green roofs, and rainwater harvesting systems (Eckart et al. 2017). These LID components are also widely known as Green Infrastructures (Zhao and Meng 2020).

In recent years, various studies have investigated the performance of LIDs or green infrastructures in urban flood control strategies. Research conducted by Palermo et al. (2020) in the catchment area of San Domenico, southern Italy, shows that applying LIDs can reduce runoff volume by 25.9%, and peak runoff by 31.4%, even though it only occupies a limited space of 30% of the total area. Similar performance is also mentioned in Li et al. (2020), where the results of the investigation on the economic area of Tianjin Airport, northern China, show that a decrease in runoff volume in the range of 4–23%, and peak discharge in the range of 4–39%, can be achieved by implementing LID practices. A higher percentage is mentioned in Movahedinia et al. (2019), which researched residential areas in District 13 of Tehran City, Iran, where applying LIDs can reduce runoff volume by 52%. Research by Guo et al. (2019) on the catchment area in Licang District, Tsingtao, Shandong, shows a 20.7–63.2% runoff reduction. Another study conducted by Ghodsi et al. (2020) stated that by applying the correct LID components with an area of less than 1 percent, a 14% reduction of runoff volume could be achieved. The effectiveness of LIDs has not only been investigated on runoff volume and discharge, but also on runoff water quality. Research conducted by Martin-Mikle et al. (2015) in downtown Oklahoma, USA, shows how LID practices can reduce nutrient and sediment loads to receiving water bodies by 16% and 17%, respectively.

Palembang City, the capital of South Sumatra province, is an area prone to flooding. One of the most significant floods occurred on 25 December 2021, when most places in the city were inundated after six hours of heavy rainfall with a depth of 168.5 mm. The depth of the flood inundation in the downtown area was reported to have reached 1.5 meters, which only receded after the rain stopped. This flood case caused two fatalities, and other impacts such as damage to houses and other building properties, and disrupted economic, social, and government activities. Figure 1 shows photo documentation of the flood inundation in Kemuning District, Palembang City. This study aims to investigate the potential level of flood reduction achieved by implementing different LID infrastructures. This study also aims to assess the ability of LIDs to restore the watershed to its natural state so that it can become an effective and sustainable flood mitigation strategy in urban areas.

Figure 1 Photo documentation of flood inundation in Kemuning District, Palembang City, on 25 December 2021 (https://palembang.tribunnews.com/).

2 Materials and method

This research was conducted in the Bendung watershed, a drainage system with a total area of 18.37 km² located in the center of Palembang City, South Sumatra province, Indonesia, as shown in Figure 2. This watershed covers six districts, including Sukarami, Kemuning, Ilir Timur I, Ilir Timur II, Kalidoni, and Ilir Barat I. The Bendung stream functions as a primary drainage channel, where its flow empties into the Musi River, one of the largest rivers in Indonesia. Therefore, the flow along the Bendung stream is still influenced by the tidal level of the Musi River.

Figure 2 Study area (Bendung watershed located in Palembang City).

The length of the main channel of the Bendung stream is about 6.3 km, with a width range of 6–12 m, 12–13.5 m, and 13.5–15 m, in the upstream, middle, and downstream reaches, respectively. The depth of the channel ranges from 1.5 m to 2.5 m. The drainage channel is mainly lined with concrete surfaces on the walls and covered by soil or sediment at the bottom due to the sedimentation process. There are three detention ponds in the upstream area, and one in the downstream area. These ponds function as temporary storages of rainfall runoff.

Land use in the Bendung drainage system is dominated by settlements and other built-up areas, such as commercial areas, offices, education, and sports facilities (primarily a golf course), and paved roads. The undeveloped areas are green-open spaces, vacant land, and water bodies. During the last ten years, there have been changes in land use where green-open space areas have been converted into residential areas, especially upstream of the watershed. Figure 3 presents the land use in the Bendung watershed in 2012 and 2022.

Figure 3 Land use within the Bendung watershed in 2012 (top), and 2022 (bottom).

The topography in the study area is mainly flat to moderately sloping, with an average slope of around 8.34%. The highest elevation is 30.85 m above mean sea level (MSL), while the lowest elevation is 1.47 m under MSL. The soil type is dominated by organosol, or histosol (peat soil), with a low permeability or poorly drained hydrologic condition. Figure 4 shows the elevation, land slope, soil type, and associated curve number (CN) maps of the study area, while Table 1 provides the land use properties. The CN values were determined based on the land use/land cover, percentage of impervious area, and hydrologic soil group (HSG), which refers to the CN table provided by Cronshey et al. (1986) and Rossman and Huber (2016a). The local climate is tropical, with a monsoonal rainfall pattern. The peak of the rainy season mainly occurs December–February, while the peak of the dry season primarily occurs in June–August. The annual rainfall in the study area for the last twenty years is about 2,591 mm/year.

Figure 4 Elevation, land slope class, soil type, and CN maps of study area.

Table 1 Land use properties in the study area.

No.	Land Use Type	Area A (Ha)	% A	% Impervious	HSG	CN (AMC II)	CN (AMC III)
1	Major Roads	42.05	2.29	100	C	98	99
2	Graveyard	28.77	1.57	50	C	79	90
3	Farmyard	0.15	0.01	25	C	74	88
4	Residential	1339.42	72.91	65	C	90	96
5	General infrastructures	0.84	0.05	85	C	94	98
6	Swamp	10.76	0.59	25	C	74	88
7	Green open space	117.05	6.37	25	C	74	88
8	Industrial facilities	9.77	0.53	72	C	91	96
9	Health facilities	16.65	0.91	85	C	94	98
10	Commercial facilities	49.10	2.67	85	C	94	98
11	Sports facilities	73.00	3.97	50	C	74	88
12	Educational facilities	68.45	3.73	85	C	94	98
13	Office facilities	31.82	1.73	85	C	94	98
14	Worship buildings/facilities	12.80	0.70	65	C	90	96
15	Vacant land	9.57	0.52	25	C	77	89
16	Water bodies	26.83	1.46	0	N/A	100	100
	Total	1837.04	100.00
	Average*			62.88	C	88.71	95.30

* The average value is proportional to each area of land use type

Six hours of heavy rainfall (24 December 2021 22:00 – 25 December 2021 03:00 UTC+7), with an accumulated depth of 168.5 mm, was recorded at the Kenten rain gauge station and is illustrated in Figure 5. The probability of this rainfall depth is estimated as a 25-year return period of maximum daily precipitation based on recorded data from 1976 to 2021. The highest rainfall intensity occurred in the 2nd hour (23:00 UTC+7) at 83 mm/hr. The 5-day antecedent rainfall (i.e., the total depth of rainfall for five days preceding the event) was 152.8 mm, calculated from recorded accumulated daily rainfall, as shown in Figure 5, which means the soils in the catchment area were saturated, and classified as antecedent moisture condition (AMC) III (Shannak et al. 2014). The observed tidal level of the Musi River is given in Figure 6. The highest and lowest tide levels on 24 December 2021 were 0.2 m (05:00 UTC+7) and 2.7 m (15:00 UTC+7), respectively, while on 25 December 2021, they were 0.3 m (06:00 UTC+7) and 2.6 m (16:00 UTC+7), respectively. The tidal water levels were used to model the drainage system as a downstream boundary condition.

Figure 5 Accumulated depth of daily rainfall recorded 1–31 December 2021 (top), and measured heavy rainfall 24 December 2021, 22.00 – 25 December 2021, 03:00 UTC+7 (bottom).

Figure 6 Observed tidal level 24 December 2021, 00:00 – 25 December 2021, 24:00 UTC+7.

The Bendung watershed was modeled in PCSWMM using an integrated 1D/2D flow model, as shown in Figure 7. PCSWMM is a next-level SWMM software with many additional features and functions, especially a 2D flood modeling module. It uses the SWMM engine, a powerful dynamic rainfall-runoff simulation model capable of simulating a range of areas, from a single uniform subcatchment to an entire city drainage system (Cahill 2012). However, unlike SWMM, which can only simulate a 1D flow regime in the channel system, PCSWMM is integrated with GIS and capable of modeling 2D flood inundations (Zhang and Valeo 2022). The 1D model components are comprised of subcatchments, nodes, links, and rain gauges. The Subcatchments component represents the catchment area, while the Nodes consist of junctions representing channel confluence and inlet points of runoff discharge from subcatchments. The Outfalls component represents the flow outlet points of a drainage system. The Links component consists of conduits representing drainage channels and orifices as flow connectors from junctions to the 2D cells. The Rain Gauge component represents rainfall as a model input. The 2D cells represent the floodplain area, a network of 2D boundary nodes with a hexagonal geometry, and a cell size of 10 m. The flow routing model used in the simulation was the dynamic wave model with a time step of 0.5 seconds. A complete explanation of the conceptual model and governing equations of the hydrology and hydraulics processes in SWMM can be found in Rossman and Huber (2016a), and Rossman (2017).

Figure 7 Schematic model of Bendung watershed in PCSWMM.

The flood analysis was initiated by modeling the flooding problem in cases of heavy rainfall on December 24–25, 2021. The hydrologic and hydraulic modeling was carried out using PCSWMM Professional 2D software, version 7.6 developed by Computational Hydraulics Inc. (CHI), which can simulate 1D and 2D flow regimes. The simulation results were then integrated with the ArcMap program to map the flood inundation. Model calibration was performed by optimizing the curve number (CN) value of the subcatchments’ property and the Manning’s roughness coefficient (n) of the conduits and floodplain area properties. The model's accuracy was examined using two evaluation parameters, the correlation coefficient (r) and Nash-Sutcliffe Efficiency coefficient (NSE), based on comparing the measured and simulated flood depths. Next, based on the calibrated model, flood mitigation simulations were carried out by implementing LIDs with various coverage areas of each land use type. Each LID’s performance was evaluated based on 2, 5, 10, and 25-year design storm simulation scenarios. The five LID infrastructures applied were bio-retention cell (BRC), rain garden (RG), permeable pavement (PP), rain barrel (RB), and recharge well (RW). The explanation of how each LID coverage area was estimated is described in Sections 2.1–2.5. The simulation results from the existing catchment (with and without LIDs) and implementation scenarios were then analyzed to investigate the performance and effectiveness of LID components in reducing flood indicators, i.e., runoff volume, peak runoff discharge, runoff peak time, runoff coefficient, flood depth, and inundation area. All stages of the analysis process in this study are illustrated in Figure 8, below.

Figure 8 Research flow chart.

2.1 Bio-retention Cell (BRC)

Bio-retention cells, or bioswales, are depressions that contain vegetation grown in an engineered soil mixture placed above a gravel drainage bed. They provide storage, infiltration, and evaporation of both direct rainfall and runoff captured from surrounding areas (Rossman and Simon 2022). The bio-retention systems consist of a surface ponding layer (0.1–0.2 m), vegetation, a soil layer (0.4–0.6 m of sandy loam), a storage layer (> 0.15 m of fine gravel), overflow structures, and an optional underdrain system (Liu et al. 2021). Typical saturated hydraulic conductivity is preferably between 100 and 300 mm/hr (ABC Waters 2018; Ballard et al. 2015). The surface area of BRCs should be 2–4% of the overall site area to be drained to prevent rapid clogging of the BRC surface (Ballard et al. 2015). The design, with an illustration of the BRCs, is given in Figure 9.

Figure 9 Design and illustration of bio-retention cell (Rossman and Simon 2022; Blecken et al. 2017).

2.2 Rain Garden (RG)

Rain gardens are a type of BRC consisting of an engineered soil layer, with no gravel bed (Rossman and Simon 2022). Basically, the typical design parameters of RG layers are like BRCs. The design, with an illustration of an RG, is given in Figure 10.

Figure 10 Design and illustration of rain garden (Rossman and Simon 2022; Zhang et al. 2020)

2.3 Permeable Pavement (PP)

Permeable pavement systems are excavated areas filled with gravel and paved with porous concrete or an asphalt mix. Block paver systems consist of impervious paver blocks placed on a sand or pea gravel bed, with a gravel storage layer below (Rossman and Simon 2022). The structure of the PP system consists of a pervious paving surface and one or more subsurface gravel lavers (Liu et al. 2021). The design, with an illustration of PP, is given in Figure 11. In this study, the PP unit is implemented in parking lots and local roads, and on several facilities, including industrial, health, commercial, educational, office, and worship facilities (Figure 12). The percentage of PP area is calculated as a fraction of the occupied area to the total area of each facility.

Figure 11 Design and illustration of permeable pavement (Rossman and Simon 2022; Cheng et al. 2019).

Figure 12 Implementation of permeable pavement on several facilities in the study area.

2.4 Rain barrel (RB)

Rain barrels or cisterns are containers that collect roof runoff during storm events and can either release or re-use the rainwater during dry periods (Rossman and Simon 2022). The most basic RB design includes a catchment surface such as a building roof, a rain gutter, a storage unit, a network of pipes connected to the catchment surface and the storage unit, and a system for transferring overflow volumes (Liu et al. 2021). The design and illustration of an RB is given in Figure 13. In this study, the typical RB unit has a 1,000 L storage capacity, with a diameter of 1.05 m and a height of 1.42 m. Each unit is assumed to capture the roof runoff from a 100 m² house plot.

Figure 13 Design and illustration of the rain barrel (Qin 2020).

2.5 Recharge Well (RW)

Recharge wells, also known as infiltration wells, are artificial holes that intercept runoff from upslope impervious areas. They provide storage volume and additional time for captured runoff to infiltrate the native soil below (Vasconcelos et al. 2019). Because the recharge well module is unavailable in PCSWMM, the infiltration trench module was used with the associated input parameters set, as shown in Table 3. The design and illustration of RW is given in Figure 14. In this study, the typical RW system consists of two cylinder holes with a diameter of 1 m and a height of 1 m (Zufrimar 2023; Mardiah et al. 2018). Each RW system is assumed to capture the roof runoff of a 100 m² house plot along with the RB unit.

Figure 14 Design and illustration of recharge well (Vasconcelos et al. 2019).

2.6 Evaluation of model accuracy

The model’s accuracy must be checked to ensure the flood problem has been modeled correctly and a calibrated model is obtained. Based on this model, the effect of the flood mitigation measure by applying the LID technique can be predicted. The two parameters of the model accuracy test used in this study were the correlation coefficient and the Nash-Sutcliffe Efficiency coefficient, which are expressed as follows:

(1)

(2)

Where:

r	=	correlation coefficient,
NSE	=	Nash-Sutcliffe Efficiency,
n	=	number of data,
	=	simulated data,
	=	observed data, and
	=	average of observed data.

The correlation coefficient ranges between 0 and 1 (0 < r ≤ 1). A value of 0 < r ≤ 0.19 means that the data has a very low correlation, 0.2 ≤ r ≤ 0.39 the data has a low correlation, 0.4 ≤ r ≤ 0.59 the data has a medium correlation, 0.6 ≤ r ≤ 0.79 the data has high correlation, and 0.8 ≤ r ≤ 1.0 the data has very high correlation. The NSE coefficient also ranges between 0 and 1 (0 < NSE ≤ 1). The coefficient of NSE < 0.36 means that the model is not qualified, 0.36 < NSE < 0.75 implies the model is qualified, and NSE > 0.75 means the model is good (Aye et al. 2017; Zhang et al. 2021).

2.7 Simulation scenarios

In this study, the possible areas that LID infrastructures can occupy are estimated proportionally based on the land use types. This is because the watershed is a densely built-up area with minimal space available for their implementation. The BRC unit is implemented in built-up areas with a coverage area of 4%, while the RG unit is placed in open space areas with a coverage area of 30%. The PP unit is also implemented in built-up areas to convert the conventional pavement of parking lots with an average coverage area of 25%. The RB and RW units are particularly placed in residential areas with coverage areas of 0.87% and 1.57%. Thus, the total area of LIDs that can be implemented is 180.22 ha or 9.81% of the catchment area. Table 2 provides details of the simulated LID implementation scenario. The initial soil moisture contents of BRC, RG, and PP units were set to be in average condition at 25%. This value was estimated from the relationship graphic between soil moisture content and soil texture for silt loam, as described in Gavrilescu (2021) and Bidkhani and Mobasheri (2018). Meanwhile, the initial saturated condition for RB and RW were set to zero, which means their initial storage would be completely empty. The percentage of impervious areas treated by BRC, PP, RB, and RW were 34.65%, 2.57%, 24.35%, and 24.35%, respectively; whereas the percentage of pervious area treated only by RG was 32.09%. These values were calculated as fractional of the total area treated by each LID unit to the total of impervious or pervious area of the catchment.

Table 2 Scenario for LID unit implementation.

LID Unit	Placement/ Land Use	Land Use Area (Ha)	% LID Unit Area (%)	LID Unit Area (Ha)	Total LID Area (Ha)	% Total LID Area (%)	Area of Each Unit (m2)	Number of Unit
BRC	Residential	937.60	4	37.50	44.47	2.42	4	111,167
	Major roads	42.05	4	1.68
	Industrial facilities	6.84	4	0.27
	Health facilities	11.65	4	0.47
	Commercial facilities	34.37	4	1.37
	Office facilities	22.27	4	0.89
	Educational facilities	47.92	4	1.92
	Worship facilities	8.96	4	0.36
RG	Green open space	117.05	30	35.12	65.65	3.57	100	6,565
	Sport facilities	73.00	30	21.90
	Graveyard	28.77	30	8.63
PP	Industrial facilities	6.84	28.69	2.80	47.26	2.57	100	4,726
	Health facilities	11.65	18.03	3.00
	Commercial facilities	34.37	26.11	12.82
	Office facilities	22.27	30.93	9.84
	Educational facilities	47.92	23.04	15.77
	Worship facilities	8.96	23.56	3.02
RB	Residential	937.60	0.87	8.12	8.12	0.44	0.87	93,760
RW	Residential	937.60	1.57	14.73	14.73	0.80	1.57	93,760
Total					180.22	9.81

BRC = bio-retention cell, RG = rain garden, PP = permeable pavement,
RB = rain barrel, RW = recharge well

The physical properties of each LID unit were adopted from several previous studies, which were then adjusted to the local conditions of the study area. Table 3 provides details of the physical parameters of the implemented LID units. In this study, the clogging factor is assumed to be zero (ignored) for all LID units, which means the flow through the LID unit layers will not be affected by clogging. A detailed explanation of each LID parameter can be found in Rossman and Simon (2022), while the conceptual model and governing equations for each LID type are well described in Rossman and Huber (2016b).

Table 3 Physical parameters of the modeled LID units.

Parameters	LID Type						References
	BRC	RG	PP	RB	RW
A. Surface							Yang et al. 2021; Randall et al. 2019; Men et al. 2020; Liang et al. 2020; Mani et al. 2019; Jato-Espino et al. 2016; Zhang et al. 2021; Mardiah et al. 2018; Vasconcelos et al. 2019; Bond et al. 2021
Berm height (mm)	300	300	100	N/A	100
Vegetation volume fraction (-)	0.1	0.5	0	N/A	0
Surface roughness (-)	0.03	0.1	0.013	N/A	0.11
Surface slope (%)	1	1	2	N/A	0
B. Pavement
Thickness (mm)	N/A	N/A	150	N/A	N/A
Void ratio	N/A	N/A	0.2	N/A	N/A
Impervious surface fraction	N/A	N/A	0.3	N/A	N/A
Permeability (mm/hr)	N/A	N/A	72	N/A	N/A
Clogging factor	N/A	N/A	0	N/A	N/A
C. Soil
Thickness (mm)	600	500	N/A	N/A	N/A
Porosity (-)	0.45	0.45	N/A	N/A	N/A
Field capacity (-)	0.11	0.12	N/A	N/A	N/A
Wilting point (-)	0.047	0.047	N/A	N/A	N/A
Conductivity (mm/hr)	155	155	N/A	N/A	N/A
Conductivity slope (-)	10	10	N/A	N/A	N/A
Suction head (mm)	50	50	N/A	N/A	N/A
D. Storage
Thickness (mm)	400	N/A	300	N/A	1100
Void ratio	0.54	N/A	0.5	N/A	0.99
Seepage rate (mm/hr)	4.5	4.5	4.5	N/A	4.5
Clogging factor	0	N/A	0	N/A	0
Barrel height (mm)	N/A	N/A	N/A	1420	N/A
E. Drain
Flow coefficient (mm/hr)	0.5	N/A	0.5	0	0.5
Flow exponent	0.5	N/A	0.5	0.5	0.5
Offset (mm)	150	N/A	100	10	1000

N/A = not applicable

Four design storms, i.e., 2-, 5-, 10-, and 25-year return periods, were used as model inputs to simulate the flood flows within the study area. The storm is distributed to a six-hour duration according to average rainfall patterns and distribution in the last ten years of recorded data. Figure 15 shows the relationship between cumulative rainfall depth and duration for each design storm. The cumulative rainfall depths for 2-, 5-, 10-, and 25-year design storms are 108.76 mm, 133.20 mm, 149.38 mm, and 169.82 mm, respectively.

Figure 15 Design storms for flood flow simulation scenarios.

3 Results and discussion

3.1 Existing conditions (flooding on 25 December 2021)

The simulation results of the existing conditions model with the input of heavy rainfall and tidal water levels from the flood case on 24 - 25 December 2021 resulted in flood inundation around the drainage system, as shown in Figure 16. The inundation locations generally occur in the upstream and middle parts of the drainage channel. This result indicates that the flooding at that time was not caused by the backwater flow effect from the Musi River, but was caused by the inability of the drainage system to manage the large volume of runoff generated by heavy rainfall, which caused the overflow. The flood extent resulting from the simulation is 192.2 ha, with inundation depths ranging from 0 – 4 m. It should be noted that inundation depths of more than 2 m occur in water bodies, including drainage channels and retention ponds. The results of this simulation are well confirmed with the flood conditions in the study area, where it is seen that the area around the main road Jl. R. Sukamto, and near the commercial area Palembang Trade Center (PTC) Mall, which was flooded during the event (Figure 17). There are 28 flood observation points available in the study area, but only 15 points are overlayed with the simulated inundation area. It indicates that other inundation areas around the study area are not traceable from the simulation results. It may occur because the inundation at other observation points is local, triggered by the water clogging and narrowing of the tertiary and secondary drainage channels, known as a bottleneck. Therefore, the evaluation of the model is only carried out using the 15 overlayed observation points.

Figure 16 Maximum flood depth map of existing condition (flood case on 25 December 2021).

Figure 17 Simulated flood inundation map. Inset: flooding around the main road Jl. R. Sukamto, and near the commercial area Palembang Trade Center (PTC) Mall.

The results of the model accuracy test show that the developed model is excellent. It is shown by the data plot between the observed and simulated flood depths, which gives a very high correlation, as shown in Figure 18. The calculated correlation coefficient of 0.987 (r = 0.987) means that the flood depths data has a very high correlation, while the Nash-Sutcliffe Efficiency coefficient of 0.948 (NSE = 0.948) means that the model performed very well. Thus, the model is suitable and can be used to predict the impact of LID implementation in reducing flooding in the study area.

Figure 18 Observed flood depth against simulated flood depth.

3.2 Flood reduction by implementing LIDs

The LIDs are implemented based on a calibrated model to simulate the impact of flood reduction. The five types of LID infrastructures are simulated with a maximum implementation area of 9.81 percent of the total catchment area. Six parameters are used as indicators to assess the impact of flood reduction by implementing LID techniques, including peak runoff discharge, total runoff volume, runoff coefficient, peak time delay, flooding area, and flood depth. All these parameters are evaluated according to four design storm scenarios.

Figure 19 provides hydrographs of total inflow at the watershed outlet (outfall) as a cumulative direct runoff from the 53 simulated subcatchments routed through the drainage channel system. Generally, it can be seen that the peak flows will decrease with the implementation of LIDs. Besides, the simulation results show a delay in the peak time of runoff, particularly on 2- and 5-year storm cases. However, in the 10- and 25-year storm design scenarios, there appears to be no delay in peak runoff time. This is because the rainfall intensities become higher in the last two scenarios, exceeding the LID unit's ability to drain. This result shows that the LID unit's capabilities are more effective in rainfall cases with lower return periods (2 and 5 years).

Figure 19 Total inflow hydrographs at the watershed outlet with and without LID implementation scenarios.

Figures 20–24, respectively, present the level of flood reduction parameters for each scenario. Generally, peak and runoff volume reduction will be lower as the return period increases. The highest peak and volume reduction rates occurred in the 2-year design storm, which are 37.7 and 35.4 percent, respectively, while the lowest reduction rates occurred in the 25-year design storm, which are 11.3 and 21.0 percent, respectively. On the other hand, the LID’s implementation can reduce the average runoff coefficient to a maximum level of 30.0 percent (0.89 to lower 0.63) and delay the peak time of runoff to 35.4 percent (1.25 hours). These results show that the implementation of LID units with a very limited occupied area of only 9.81% can significantly impact the reduction of flood levels. However, the level of flood extent reduction in this study will vary as the return period increases, with an average level of 17.91 percent. This result possibly happens because the topographic conditions of the catchment area greatly influence the distribution of flood inundation.

Figure 20 Runoff volume reduction for each simulation scenario.

Figure 21 Runoff peak reduction for each simulation scenario.

Figure 22 Runoff coefficient reduction for each simulation scenario.

Figure 23 Runoff peak time delay for each simulation scenario.

Figure 24 Flood extent reduction for each simulation scenario.

The ability of the LID to reduce runoff and flooding indicates that each LID component can imitate and restore watershed hydrological conditions to their natural state. It is shown by the decrease in the runoff coefficient (Figure 22). The runoff coefficient ranges between 0 and 1 (0 ≤ C ≤ 1), where C = 0 means there is no runoff, while C = 1 means all rainfall becomes runoff (Zhang et al. 2022). In the figure, the average runoff coefficient from the existing condition of the study area without LIDs is 0.89 – 0.93. It indicates that almost all the rainfall in the study area will be converted into a runoff. However, with the implementation of the LIDs, the runoff coefficient can be reduced to 0.63 – 0.75.    

LID implementation can reduce the area and depth of inundation. The ability of the LID structures to hold and store rainfall runoff can reduce the flow discharge that enters the drainage channels, which has implications for lowering overflow. Besides that, the LIDs can also help reduce the imbalance between runoff increasing and limited drainage channel capacity, which tends to be static, and even decreases due to sedimentation. Figure 25 shows the maps of the inundation areas for each simulation scenario. The inundation areas will increase as the rainfall return period increases. This result is very reasonable considering that the volume of runoff received by the drainage system will be greater, causing more significant overflow discharge. The inundation areas in the 2-, 5-, 10-, and 25-year storm scenarios are 130.3, 148.6, 159.8, and 180.0 hectares, respectively. However, although the increasing inundation areas occurred evenly along the drainage system, it predominantly occurred in the middle stream (Figure 26). This happens because the middle stream is a confluence of the two main upstream channels, so the flow accumulation at that location will increase significantly, causing a wider inundation area.

Figure 25 Flood extent maps for each design storm.

Figure 26 Flood extent maps zoomed to the location of the upstream confluence.

Figure 27 presents a flood depth plot from 15 sampling points for each simulation scenario. Generally, the number of flood spots will increase as the rainfall return period increases. In addition, the inundation depths will also increase with higher return periods and vice versa, caused by the impact of expanding the flood inundation area. Based on these results, it can be said that the implementation of LIDs successfully reduces the number and magnitude of flood depths and flood extents in the study area.

Figure 27 Flood depths of fifteen sampling points.

3.3 Flood mitigation effectiveness

The effectiveness of flood mitigation by implementing LIDs can be associated with their ability to reduce flood indicators. The simulation results show that the lower the return period of design rainfall, the greater the level of flood reduction that can be achieved. In other words, the performance of LIDs will be more effective for lower rainfall magnitude. The increase in flood reduction effectiveness tends to be in the nonlinear form (power trendline) to the decrease of return period, as shown in Figure 28. Applying combined LIDs with a maximum occupied area of 9.81 percent of the total study area can achieve more than 30% flood reduction. It confirms that the implementation of LIDs is very effective and promising in urban flood mitigation planning. This result also confirms a study by Movahedinia et al. (2019) that shows combining various LID components can provide more effective results in reducing runoff volume and peak flood discharge.

Figure 28 Relationship between the reduction level of flood indicators and the return period of design rainfall in the study area.

3.4 Discussion

The phenomenon of flooding can be caused by the inability of the drainage system to manage flood flows. In urban areas, increased surface runoff due to raised built-up and impervious areas can further increase the discharge load into canals and rivers. On the other hand, the capacity improvement of the drainage channels tends to be stagnant and even decreases due to sedimentation. Structural measures by channel widening are sometimes difficult or impossible to carry out in dense areas, requiring land acquisition at a high cost and potentially triggering social conflict. The LID technique has become one of the best alternatives for flood mitigation, especially in dense urban areas with limited space for its implementation (Azkarini et al. 2019).

The ability of LIDs to reduce flooding has been demonstrated in this study. The combination of five LID infrastructures, namely BRC, RG, PP, RB, and RW, can significantly reduce flood parameters such as runoff volume, peak runoff discharge, runoff coefficient, peak time, flood depth, and flooding area. The effectiveness of more than 30 percent flood reduction can be achieved with the LID occupied area only 9.81 percent of the total watershed area. In addition, the results of this study also confirm LIDs’ ability to mimic and restore watershed hydrological conditions to their natural state, where the runoff coefficient can be reduced to 0.63 – 0.75. Reducing the volume of surface runoff by implementing LIDs can help reduce the flow load within the drainage system in the study area. However, its effectiveness will decrease in very heavy or extreme rainfall cases. Therefore, to maximize the structural measure of urban flood control, it seems that it must be followed by increasing the capacity of the drainage system.

Further research is needed by optimizing the combination of LID infrastructures to obtain a broader picture of how each LID unit affects flood reduction. Besides that, further research should investigate how the LIDs become feasible to be constructed in urban areas with limited financial capacity. Because the costs required for the construction of each LID type are varied and can be high, it becomes a constraint that needs to be considered in its implementation (Putri et al. 2023). The level of runoff reduction also needs to be tested against models with long and continuous input data series (monthly and yearly) from hydrological variables such as rainfall, evapotranspiration, infiltration, and others (Guo et al. 2018). In addition, the clogging factor also needs to be considered so that the LID performance can be monitored over time (Mani et al. 2019; Hu et al. 2018). Furthermore, further research that investigates the impact of LID implementation on water and air quality remediation and improving the local microclimate quality should also be addressed (Kasprzyk et al. 2022). This is because implementing LIDs can control not only the problem of runoff quantity, but also its quality, which has a broad impact on the surrounding environment.

4 Conclusion

Based on the results and discussion above, it can be concluded that implementing the LID technique can significantly reduce surface runoff, which can reduce flood parameters. It can be achieved in an urban area even with very limited or available occupied areas. The ability of LID components to hold and store runoff flows will significantly help to minimize the drainage system load. This study has demonstrated that with only about 9.81 percent of the area occupied by LIDs, a flood reduction of more than 30% can be achieved. Besides, implementing LIDs can help restore the watershed hydrological condition to its natural state, as indicated by the decrease in the runoff coefficient. Thus, the LID technique as a sustainable urban drainage system must be widely promoted in many urban areas, especially in developed countries like Indonesia. This study can be used as a reference to the government and local authorities to create policies and regulations to establish a sustainable flood mitigation measure in Palembang City.

This study only focuses on flood mitigation by LIDs within short-term periods (flooding-event flows). Future studies should focus on investigating the performance of LIDs for long-term and continuous simulations. Rainfalls and other climate factors such as evapotranspiration, temperature, and solar radiation simultaneously with groundwater flows can be included in the LID model to provide more comprehensive results about the water balance in urban watershed areas.

Acknowledgments

We want to thank the Indonesia Endowment Fund for Education Agency (LPDP) for funding this research and the Center for Environmental Studies, Universitas Gadjah Mada, for supporting this research. Also, we thank Computational Hydraulics International (CHI) and ESRI Indonesia for providing the license and software of PCSWMM Professional 2D and ArcGIS Desktop to support this research through a university collaboration program. The authors also thank all contributors and local authorities who have assisted in providing the data.

References

1. ABC Waters. 2018. Condensed Booklet on Engineering Procedures for ABC Waters Design Features. 2018 Edition. PUB Singapore’s National Water Agency: Singapore.
2. Arahuetes, A., and J.O. Cantos. 2019. “The Potential of Sustainable Urban Drainage Systems (SuDS) as an Adaptive Strategy to Climate Change in the Spanish Mediterranean.” International Journal of Environmental Studies 76 (5): 764–79. https://doi.org/10.1080/00207233.2019.1634927
3. Aye, P.P., S. Koontanakulvong, and T.T. Long. 2017. “Estimation of Groundwater Flow Budget in the Upper Central Plain, Thailand from Regional Groundwater Model.” Internet Journal of Society for Social Management Systems 11 (1): 90–100. https://ssms.jp/img/files/2019/04/sms17_3216.pdf
4. Azkarini, L., E. Anggraheni, and D. Sutjiningsih. 2019. “The Influence of Low Impact Development-Best Management Practices Implementation on Surface Runoff Reduction: A Case Study in Universitas Indonesia Catchment Area.” MATEC Web of Conferences 276: 04007. https://doi.org/10.1051/matecconf/201927604007
5. Ballard, B.W., S. Wilson, H. Udale-Clark, S. Illman, T. Scott, R. Ashley, and R. Kellagher. 2015. The SuDS Manual. CIRIA C753: London, UK.
6. Bidkhani, N.O.G., and M.R. Mobasheri. 2018. “Influence of Soil Texture on the Estimation of Bare Soil Moisture Content Using MODIS Images.” European Journal of Remote Sensing 51 (1): 911–920. https://doi.org/10.1080/22797254.2018.1514986
7. Blecken, G-T., W.F. Hunt, A.M. Al-Rubaei, M. Viklander, and W.G. Lord. 2017. “Stormwater Control Measure (SCM) Maintenance Considerations to Ensure Designed Functionality.” Urban Water Journal 14 (3): 278–290. https://doi.org/10.1080/1573062X.2015.1111913
8. Bond, J., E. Batchabani, M. Fuamba, D. Courchesne, and G. Trudel. 2021. “Modeling a Bioretention Basin and Vegetated Swale with a Trapezoidal Cross Section Using SWMM LID Controls.” Journal of Water Management Modeling 29 (C474): 1–15. https://doi.org/10.14796/JWMM.C474
9. Cahill, T.H. 2012. Low Impact Development and Sustainable Stormwater Management: Cahill/Sustainable Stormwater. John Wiley & Sons, Inc.: Hoboken, NJ, USA. https://doi.org/10.1002/9781118202456
10. Chan, F.K.S., J.A. Griffiths, D. Higgitt, S. Xu, F. Zhu, Y-T. Tang, Y. Xu, and C.R. Thorne. 2018. “‘Sponge City’ in China—A Breakthrough of Planning and Flood Risk Management in the Urban Context.” Land Use Policy 76: 772–78. https://doi.org/10.1016/j.landusepol.2018.03.005
11. Chen, Y., H.W. Samuelson, and Z. Tong. 2016. “Integrated Design Workflow and a New Tool for Urban Rainwater Management.” Journal of Environmental Management 180: 45–51. https://doi.org/10.1016/j.jenvman.2016.04.059
12. Cheng, Y-Y., S-L. Lo, C-C. Ho, J-Y. Lin, and S.L. Yu. 2019. “Field Testing of Porous Pavement Performance on Runoff and Temperature Control in Taipei City.” Water 11 (12): 2635. https://doi.org/10.3390/w11122635
13. Cronshey, R., R.H. McCuen, N. Miller, W. Rawls, S. Robbins, and D. Woodward. 1986. Urban Hydrology for Small Watersheds. Second Edition. TR-55. U. S. Department of Agriculture (USDA).
14. Eckart, K., Z. McPhee, and T. Bolisetti. 2017. “Performance and Implementation of Low Impact Development – A Review.” Science of The Total Environment 607–608: 413–32. https://doi.org/10.1016/j.scitotenv.2017.06.254
15. Fletcher, T.D., W. Shuster, W.F. Hunt, R. Ashley, D. Butler, S. Arthur, S. Trowsdale, et al. 2015. “SUDS, LID, BMPs, WSUD and More – The Evolution and Application of Terminology Surrounding Urban Drainage.” Urban Water Journal 12 (7): 525–542. https://doi.org/10.1080/1573062X.2014.916314
16. Gavrilescu, M. 2021. “Water, Soil, and Plants Interactions in a Threatened Environment.” Water 13 (19): 2746. https://doi.org/10.3390/w13192746
17. Ghodsi, S.H., Z. Zahmatkesh, E. Goharian, R. Kerachian, and Z. Zhu. 2020. “Optimal Design of Low Impact Development Practices in Response to Climate Change.” Journal of Hydrology 580: 124266. https://doi.org/10.1016/j.jhydrol.2019.124266
18. Goh, X.P., M. Radhakrishnan, C. Zevenbergen, and A. Pathirana. 2017. “Effectiveness of Runoff Control Legislation and Active, Beautiful, Clean (ABC) Waters Design Features in Singapore.” Water 9 (8): 627. https://doi.org/10.3390/w9080627
19. Guo, C., J. Li, H. Li, B. Zhang, M. Ma, and F. Li. 2018. “Seven-Year Running Effect Evaluation and Fate Analysis of Rain Gardens in Xi’an, Northwest China.” Water 10 (7): 944. https://doi.org/10.3390/w10070944
20. Guo, X., P. Du, D. Zhao, and M. Li. 2019. “Modelling Low Impact Development in Watersheds Using the Storm Water Management Model.” Urban Water Journal 16 (2): 146–155. https://doi.org/10.1080/1573062X.2019.1637440
21. Herath, H.M.M.S.D., T. Fujino, and M.D.H.J. Senavirathna. 2023. “A Review of Emerging Scientific Discussions on Green Infrastructure (GI)-Prospects towards Effective Use of Urban Flood Plains.” Sustainability 15 (2): 1227. https://doi.org/10.3390/su15021227
22. Hu, M., X. Zhang, Y. Siu, Y. Li, K. Tanaka, H. Yang, and Y. Xu. 2018. “Flood Mitigation by Permeable Pavements in Chinese Sponge City Construction.” Water 10 (2): 172.
    https://doi.org/10.3390/w10020172
23. Jato-Espino, D., S.M. Charlesworth, J. Bayon, and F. Warwick. 2016. “Rainfall–Runoff Simulations to Assess the Potential of SuDS for Mitigating Flooding in Highly Urbanized Catchments.” International Journal of Environmental Research and Public Health 13 (1): 149. https://doi.org/10.3390/ijerph13010149
24. Kasprzyk, M., W. Szpakowski, E. Poznańska, F.C. Boogaard, K. Bobkowska, and M. Gajewska. 2022. “Technical Solutions and Benefits of Introducing Rain Gardens – Gdańsk Case Study.” Science of The Total Environment 835: 155487. https://doi.org/10.1016/j.scitotenv.2022.155487
25. Li, Y., J.J. Huang, M. Hu, H. Yang, and K. Tanaka. 2020. “Design of Low Impact Development in the Urban Context Considering Hydrological Performance and Life‐cycle Cost.” Journal of Flood Risk Management 13 (3): e12625. https://doi.org/10.1111/jfr3.12625
26. Liang, C., X. Zhang, J. Xia, J. Xu, and D. She. 2020. “The Effect of Sponge City Construction for Reducing Directly Connected Impervious Areas on Hydrological Responses at the Urban Catchment Scale.” Water 12 (4): 1163. https://doi.org/10.3390/w12041163
27. Liu, T., Y. Lawluvy, Y. Shi, and P-S. Yap. 2021. “Low Impact Development (LID) Practices: A Review on Recent Developments, Challenges and Prospects.” Water, Air, & Soil Pollution 232: 344. https://doi.org/10.1007/s11270-021-05262-5
28. Mani, M., O. Bozorg-Haddad, and H.A. Loáiciga. 2019. “A New Framework for the Optimal Management of Urban Runoff with Low-Impact Development Stormwater Control Measures Considering Service-Performance Reduction.” Journal of Hydroinformatics 21 (5): 727–744. https://doi.org/10.2166/hydro.2019.126
29. Mardiah, A.M., C.N. Ainy, M. Bagus, and D. Harlan. 2018. “Study on the Effectiveness of Infiltration Wells to Reduce Excess Surface Run Off In ITB.” MATEC Web of Conferences 147: 03008. https://doi.org/10.1051/matecconf/201814703008
30. Martin-Mikle, C.J., K.M. de Beurs, J.P. Julian, and P.M. Mayer. 2015. “Identifying Priority Sites for Low Impact Development (LID) in a Mixed-Use Watershed.” Landscape and Urban Planning 140: 29–41. https://doi.org/10.1016/j.landurbplan.2015.04.002
31. Men, H., H.L., W. Jiang, and D. Xu. 2020. “Mathematical Optimization Method of Low-Impact Development Layout in the Sponge City.” Mathematical Problems in Engineering 1: 1–17. https://doi.org/10.1155/2020/6734081
32. Montazerolhodjah, M. 2019. “Urban Environments Sustainable Development through Low Impact Approaches.” Progress in Industrial Ecology – An International Journal 13 (1): 16–28.
33. Movahedinia, M., J.M.V. Samani, F. Barakhasi, S. Taghvaeian, and R. Stepanian. 2019. “Simulating the Effects of Low Impact Development Approaches on Urban Flooding: A Case Study from Tehran, Iran.” Water Science and Technology 80 (8): 1591–1600. https://doi.org/10.2166/wst.2019.412
34. Nowogoński, I. 2020. “Low Impact Development Modeling to Manage Urban Stormwater Runoff: Case Study of Gorzów Wielkopolski.” Journal of Environmental Engineering and Landscape Management 28 (3): 105–115. https://doi.org/10.3846/jeelm.2020.12670
35. Palermo, S.A., V.C. Talarico, and M. Turco. 2020. “On the LID Systems Effectiveness for Urban Stormwater Management: Case Study in Southern Italy.” IOP Conference Series: Earth and Environmental Science 410: 012012. https://doi.org/10.1088/1755-1315/410/1/012012
36. Putri, F.K., E. Hidayah, and M.F. Ma’ruf. 2023. “Enhancing Stormwater Management with Low Impact Development (LID): A Review of the Rain Barrel, Bioretention, and Permeable Pavement Applicability in Indonesia.” Water Science & Technology 87 (9): 2345–2361. https://doi.org/10.2166/wst.2023.095
37. Qin, Y. 2020. “Urban Flooding Mitigation Techniques: A Systematic Review and Future Studies.” Water 12 (12): 3579. https://doi.org/10.3390/w12123579
38. Randall, M., F. Sun, Y. Zhang, and M.B. Jensen. 2019. “Evaluating Sponge City Volume Capture Ratio at the Catchment Scale Using SWMM.” Journal of Environmental Management 246: 745–757. https://doi.org/10.1016/j.jenvman.2019.05.134
39. Rossman, L.A. 2017. Storm Water Management Model Reference Manual Volume II – Hydraulics. EPA/600/R-17/111. U.S. Environmental Protection Agency, Washington, DC. https://cfpub.epa.gov/si/si_public_record_report.cfm?Lab=NRMRL&dirEntryId=337162
40. Rossman, L.A., and W.C. Huber. 2016a. Storm Water Management Model Reference Manual Volume I, Hydrology (Revised). EPA/600/R-15/162A. U.S. Environmental Protection Agency, Washington, DC. https://cfpub.epa.gov/si/si_public_record_report.cfm?Lab=NRMRL&dirEntryId=309346
41. Rossman, L.A., and W.C. Huber. 2016b. Storm Water Management Model Reference Manual Volume III – Water Quality. EPA/600/R-16/093. United States Environmental Protection Agency, Washington, D.C. https://cfpub.epa.gov/si/si_public_record_report.cfm?Lab=NRMRL&dirEntryId=327450
42. Rossman, L.A., and M. Simon. 2022. Storm Water Management Model User’s Manual Version 5.2. EPA/600/R-22/030. United States Environmental Protection Agency, Washington D.C. https://www.epa.gov/system/files/documents/2022-04/swmm-users-manual-version-5.2.pdf
43. Shannak, S.A., F.H. Jaber, and B.J. Lesikar. 2014. “Modeling the Effect of Cistern Size, Soil Type, and Irrigation Scheduling on Rainwater Harvesting as a Stormwater Control Measure.” Water Resources Management 28 (12): 4219–4235. https://doi.org/10.1007/s11269-014-0740-x
44. Tan, K.M., W.K. Seow, C.L. Wang, H.J. Kew, and S.B. Parasuraman. 2019. “Evaluation of Performance of Active, Beautiful and Clean (ABC) on Stormwater Runoff Management Using MIKE URBAN: A Case Study in a Residential Estate in Singapore.” Urban Water Journal 16 (2): 156–162. https://doi.org/10.1080/1573062X.2019.1634744
45. Vasconcelos, A.F., T.S. Ferreira, M.F. N. dos Santos, and A.P. Barbassa. 2019. “Modeling Infiltration Wells in SWMM and Comparing Its Performance with a Real-Scale Well.” In New Trends in Urban Drainage Modelling, ed. G. Mannina, Green Energy and Technology, 424–28. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-319-99867-1_72
46. Yang, Y., J. Li, Q. Huang, J. Xia, J. Li, D. Liu, and Q. Tan. 2021. “Performance Assessment of Sponge City Infrastructure on Stormwater Outflows Using Isochrone and SWMM Models.” Journal of Hydrology 597: 126151. https://doi.org/10.1016/j.jhydrol.2021.126151
47. Yin, D., C. Xu, H. Jia, Y. Yang, C. Sun, Q. Wang, and S. Liu. 2022. “Sponge City Practices in China: From Pilot Exploration to Systemic Demonstration.” Water 14 (10): 1531. https://doi.org/10.3390/w14101531
48. Zhang, L., Z. Ye, and S. Shibata. 2020. “Assessment of Rain Garden Effects for the Management of Urban Storm Runoff in Japan.” Sustainability 12 (23): 9982. https://doi.org/10.3390/su12239982
49. Zhang, Y., H. Xu, H. Liu, and B. Zhou. 2021. “The Application of Low Impact Development Facility Chain on Storm Rainfall Control: A Case Study in Shenzhen, China.” Water 13 (23): 3375. https://doi.org/10.3390/w13233375
50. Zhang, Z., W. Hu, W. Wang, J. Zhou, D. Liu, X. Qi, and X. Zhao. 2022. “The Hydrological Effect and Uncertainty Assessment by Runoff Indicators Based on SWMM for Various LID Facilities.” Journal of Hydrology 613A: 128418. https://doi.org/10.1016/j.jhydrol.2022.128418
51. Zhang, Z., and C. Valeo. 2022. “Verification of PCSWMM’s LID Processes and Their Scalability over Time and Space.” Frontiers in Water 4: 1058883. https://doi.org/10.3389/frwa.2022.1058883
52. Zhao, G., and D.Z. Meng. 2020. “Research Progress of Low Impact Development Technology.” IOP Conference Series: Earth and Environmental Science 474: 052033. https://doi.org/10.1088/1755-1315/474/5/052033
53. Zufrimar, E.Z. 2023. “Infiltration Wells for Various House Types in the Kuranji Catchment Area, Padang City.” In AIP Conference Proceedings 2691:030006. https://doi.org/10.1063/5.0114963