Abstract

To assess the urban waterlogging mitigation effectiveness on low impact development (LID) in semi-mountainous regions, the Storm Water Management Model (SWMM) of a semi-mountainous region combined with GIS was generalized. The SWMM was calibrated and validated through maximum seeper depth of the checkpoints, and various LID scenarios have been designed according to local conditions. The discharge processes of outlets, surface runoff, peak flow and peak time were analyzed in different scenarios. The results show that: all the flow processes of outlets in the LID scenario are gentler than that in the status quo scenario, and the effectiveness of LIDs in semi-mountainous regions are different from that in plain regions because of the slope influence; in semi-mountainous regions, the LID effectiveness on surface runoff reduction decreases with the increase in rainfall return period or the extension of rainfall duration, but remains almost unchanged with the increase in rainfall peak coefficient; the LID effectiveness on control peak flow reduction is not remarkable with the change in rainfall characteristics, and the LID effectiveness on peak time delay is poor. This research can provide decision support for regional small-scale measures of urban waterlogging mitigation and reduction in semi-mountainous regions.

HIGHLIGHTS

  • The SWMM of a typical semi-mountainous region was established utilizing ArcGIS and the surveyed maximum seeper depth.

  • The numbers, properties, and locations of LIDs were designed based on the local characteristics.

  • The results indicated that LIDs effectiveness is related to slope in the study area.

  • This research can provide technique support to resilient cities design in semi-mountainous regions.

INTRODUCTION

Recently, rapid urbanization in developing countries has caused dramatic changes in regional land characteristics including sharp increase of impervious surfaces, and has led to gradual changes in regional hydrologic cycles (Gunn et al. 2012), which are usually manifested in the acceleration of runoff velocity, increase of runoff discharge and decrease of runoff duration. In addition, the infiltration volumes of rainwater, groundwater recharge, base flow, and evapotranspiration have decreased in urbanized regions (Tang et al. 2005; Jacobson 2011; Sun et al. 2011). The above changes in urban hydrologic processes have resulted in frequent urban floods, water environment pollution, and water resource shortages (Sun et al. 2011; Wang et al. 2017). In order to solve the above problems, China proposed a sponge city construction plan in 2014, attempting to find ecologically suitable alternatives to mitigate water-related problems such as urban floods (MHURD 2014; Hu et al. 2019). In the sponge city construction plan, low impact development practices (LIDs), proposed in the 1990s by the United States Environmental Protection Agency (USEPA) to restore the urban hydrologic cycle to a natural or near-natural status, is a critical component (Mei et al. 2018).

With the rapid improvement of computer techniques, the utilization of simulation models for urban flood research is becoming more frequent. The storm water management model (SWMM) is one of the commonly used drainage system models for simulating and evaluating urban waterlogging. Through the LID module of SWMM or other models, a variety of different LID assessments can be simulated and quantified, which provides a good foundation for the related research (Rossman 2014; Bai et al. 2019). Numerous studies have confirmed the effectiveness of LIDs for urban waterlogging (Ahiablame et al. 2012; Fletcher et al. 2015; Hu et al. 2019). The studies indicated that through LIDs, the flood peak can be decreased, the runoff coefficient can be reduced, and the hydrologic response time can be prolonged (Chen et al. 2016; Soulis et al. 2017). LID utilizes bioretention, permeable pavements, vegetative swales, and other technical facilities to mitigate urban flood and improve the sustainable use of water resources, in order to maintain and protect the natural hydrologic function of application regions (Traver & Ebrahimian 2017). At present, LIDs are regarded as a promising and effective strategy for urban waterlogging reduction, water pollution mitigation, and the improvement of urban drainage systems (Maniquiz-Redillas & Kim 2016; Eckart et al. 2017).

The hydrologic performance of LIDs has been studied at different scales, including laboratory scale, pilot scale, and in-situ full scale (Qin et al. 2013). For instance, Alfredo et al. (2010) found that green roofs can delay and prolong roof discharge, and reduce its peak by 30–78% compared with a standard roof in the laboratory. Some studies indicted that the total runoff and flood peak of the catchment decreased in the LID scenario compared with those in the traditional scenario, and the surface runoff process through LIDs was similar to that in the undeveloped status (Dietz & Clausen 2008; Bedan & Clausen 2009). Hu et al. (2017) evaluated LID performance on mitigating flood waterlogging at a watershed scale in China and found that LIDs can effectively reduce the waterlogging area at the watershed scale, especially in the high risk regions.

Moreover, the performances of single or combined LIDs in controlling urban waterlogging have been studied and focused on. Li et al. (2018) investigated the operational effectiveness of bioretention on different tanks through tests and simulations, and found that return periods, media layer thickness, and solute concentration were the significant sensitivity parameters. Rodríguez-Rojas et al. (2018) analyzed the hydrologic performances of three types of permeable pavements for reducing flood risk such as delaying catchment responses and slowing flow velocities and found them significant; additionally, these pavements can also improve the local circumstances of cities and safety for pedestrians and vehicle traffic. Other studies (Liu et al. 2015; Palla & Gnecco 2015; Zhu & Chen 2017) found that various BMPs/LID combinations can also control runoff and improve water quality effectively.

Further studies indicated that the LID performances in urban waterlogging control are significantly different in various storms. Hood et al. (2007) compared the runoff volume, peak discharge, and runoff coefficient of LID mode with those of traditional development mode in a typical town of Waterford and illustrated that the LID effects on runoff reduction of smaller storms with shorter durations were greater. Due to complicated climate circumstance and fast urbanization, many related case studies have focused on China. For instance, Mei et al. (2018) designed 15 scenarios for LID construction under six designed rainfall events for integrated assessments of LIDs and the assessment results of Liangshuihe watershed in Beijing indicated that the reduction rates of flood-related indices decreased with the increase of rainfall return periods. Bai et al. (2019) found that the runoff reduction of LIDs is more effective on shorter and heavier rainfall events in Suqian City, Jiangsu Province. Hu et al. (2019) evaluated the hydrologic response of LID on different rainfall duration and quantity in Nanjing, and the results indicated that the flood mitigation performance decreased sharply with quantity increase, but slowly with duration increase. Zheng et al. (2019) simulated the runoff control effects of seven LID scenarios on different rainfall characteristics and found that LID scenarios are more effective on flood control in smaller return periods and longer duration storms, and the peak locations of rainfall process have no significant influence on runoff generation.

In brief, several studies focused on LID effectiveness in different rainfall scenarios widely recognized that LIDs are effective on the reduction of runoff volume and peak flow. However, most of the related studies are applied in plains, with only a few applied in mountainous or semi-mountainous regions (Luan et al. 2017; Wang et al. 2019). Especially, few studies focused on the hydrologic effectiveness of LID in different rainfall scenarios. Due to the various topographies and complex circumstances in mountainous or semi-mountainous areas, it is necessary to illustrate the hydrologic performance of LIDs on urban waterlogging mitigation in different rainfall processes.

This study assessed the LID effectiveness on urban waterlogging control during different rainfall events in a typical semi-mountainous area of China, through the calibrated and validated SWMM, in order to enrich the LID studies in different topographic areas, and supply the technologic support for urban storm management in semi-mountainous areas. In this paper, ‘urban waterlogging’ refers to the phenomenon that the accumulation of water in an urban region due to heavy or continuous rainfall exceeds the capacity of urban drainage systems.

MATERIALS AND METHODS

Study area

The study area is located in the central district of Lincheng County (CDLC) in Xingtai City, Southwest Hebei Province, China (Figure 1). Lincheng County is located in a warm and semi-humid continental monsoon climate zone with an annual average precipitation of 605 mm, and the precipitation in this area is mainly concentrated from June to August. The total area is 11.8 km2.

Figure 1

Location and land use map of the study area.

Figure 1

Location and land use map of the study area.

The study area is bordered by Xiaohuai River in the north and Zhi River in the west and is crossed by the South-to-North Water Diversion in the middle. According to the land use distributions (shown in Figure 1), the study area was divided into west, middle, and east areas, which are 14.8%, 35.4%, and 49.8% of the total area, respectively. Each part is an independent drainage system connected to one outlet. The west area is mainly an old residential zone; the middle area is a new residential zone for public and commercial service; and the east area is mainly an industrial zone.

Lincheng County, as a typical mountainous area, is located at the eastern foothills of the Taihang Mountains. The middle and low mountains are in the west area, hills are in the middle area, and plains are in the east area. Owing to its steep topography and large slope, large urban waterlogging is frequent when it rains heavily.

Data collection

In this study, various data were needed, mainly including drainage system data, land use data, topographic data, precipitation data, and the investigated maximum seeper depth (MSD) of the precipitation process. The drainage system data, land use data (Figure 1), and the elevations of drainage pipeline nodes were directly obtained from the ‘Detailed Regulatory Plan for the Central District of Lincheng County (2013–2030)’, which was provided by Norendar International Ltd in Shijiazhuang, Hebei, China.

Topographic data

Through ArcGIS, the topographic data (Figure 2) were generated from the elevations of drainage pipeline nodes and the digital elevation model (DEM), which was provided by Hebei Bureau of Geoinformation, in Shijiazhuang, Hebei, China, and the elevation and slope distributions of the study area are illustrated in Figure 2(a) and 2(b), respectively. Overall, the elevation in the west area gradually decreases from southeast to northeast, then to southwest; the highest land is approximately 120 m and the lowest is approximately 90 m. The elevation in the middle area is much higher, and decreases gradually from southwest to northeast; the highest is approximately 132 m and the lowest is approximately 96 m. The elevation in the east area is relatively low and decreases gradually from southwest to northeast, where the lowest land is only 84.6 m. The change of slope distribution corresponds with the elevation distribution. Most of the slopes in the west area are between 1 and 3°, and a few slopes are less than 1°. The slopes in the middle area fluctuate sharply from 12° in the steep zones to less than 1° in the gentle zones. Mst of the terrain in the east area slopes gently from 1.7° to 0.5°, and several slopes are even below 0.5°.

Figure 2

Distributions of elevation (a) and slope (b) in the study area. (a) Elevation (m), (b) Slope (°).

Figure 2

Distributions of elevation (a) and slope (b) in the study area. (a) Elevation (m), (b) Slope (°).

Therefore, this study area is a typical semi-mountainous region according to the above illustration. Because LIDs are various, such as swales, permeable pavements, and green roofs, the topographic adaptability of each sort is different. These related analyses on elevation and slope are crucial to integrate effective LIDs in the whole area for mitigating urban waterlogging.

Precipitation data

The precipitation data were provided by Hydrology and Water Resources Investigation Bureau of Hebei, China. Two typical processes (Figure 3) which can cause the surface overflow were selected. One typical process was on July 19, 2016 (hereafter, ‘7.19’ storm), shown in Figure 3, and is the largest one in Lincheng County since August 1996. It began at 4:00 on July 19 and ended at 20:15 on July 20, of which the peak time occurred at 9:50 on July 19. The total precipitation reached 220 mm, which caused serious urban waterlogging and disasters. This storm process is a typical heavy one, selected to calibrate the parameters.

Figure 3

The ‘7.19’ and ‘5.22’ precipitation processes.

Figure 3

The ‘7.19’ and ‘5.22’ precipitation processes.

The other process was on May 22, 2017 (hereafter, ‘5.22’ storm), shown in Figure 3. It began at 14:30 on May 22 and ended at 1:15 on May 21, of which the peak time occurred at 16:10 on May 22. It is a typical moderate rain event with a total precipitation of 37.4 mm, which was selected for model validation.

MSD data

MSD data on different positions of two storm processes were collected through manual investigation and survey (Figure 4). According to the surveyed low-lying and urban waterlogging distributions in the sub-catchments, five positions (Figure 4) were selected: 1# is located at the northeast corner of the east area, at the intersection of Weiyi Road and Jingsan Street; 2# is located in the middle of the middle area, at the intersection of Lincheng Avenue and Linquan East Road; 3# is located in the middle of the middle area, at the intersection of Linquan East Road and Youyi Street; 4# is located in the middle of the west area, at the intersection of Wenhua Street and Changzheng Road; and 5# is located in the northwest corner of the west area, at the intersection of Xinglin Street and Xingfu Road. All of these positions often experience serious flooding and waterlogging. The accurate MSDs of ‘7.19’ and ‘5.22’ storm processes in each position were collected through detailed and repeated investigation and survey.

Figure 4

Distribution of the survey positions and generalized results of the drainage system.

Figure 4

Distribution of the survey positions and generalized results of the drainage system.

Hydrological model setup

Model selection

The USEPA SWMM (Rossman 2014; Kong et al. 2017) was selected in this study. The SWMM as a hydrological model is frequently used for urban water quantity simulation in different precipitation events based on the Saint-Venant system of equations (Equations (1) and (2)) (Tan et al. 2008; Zhu et al. 2019). The urban catchment consists of different sub-catchments and each sub-catchment is the basic computation unit that receives rainfall and generates surface runoff, infiltration, etc. (Palla & Gnecco 2015). The simulation method of the rainfall-runoff process in SWMM is nonlinear reservoir and the calculation methods of infiltration process include Horton, Green–Ampt, and Soil Conservation Service Curve Number (SCS-CN) models. In previous studies on SWMM for urban areas in a semi-drought region, the Horton method was successfully applied to calculate the infiltration process (Luan et al. 2017; Mei et al. 2018). Thus, this study also chooses the Horton model, which was calculated through Equation (3), for calculating small watersheds. The methods of flow routing consist of the steady flow, kinematic wave, dynamic wave methods, and the dynamic wave routing method is used in this paper, which by solving the complete one-dimensional Saint-Venant equations, the complicated motion problem of water flow in closed pipe sections can be effectively solved compared to the other methods. A reasonable flow calculation can be achieved used the dynamic wave routing method whether it is at the connection of upstream and downstream pipe sections or between pipe sections with backwater and countercurrent (Rossman 2014). In addition, the SWMM can also simulate other physical processes such as water quality, snow-melt, evaporation, etc. (Rossman 2014).
formula
(1)
formula
(2)
where Q is the quantity of flow (m3/s), A is the area of the water cross-section (m2), H is the water depth (m), g is the acceleration of gravity (m/s2), and Sf is the friction ratio.
formula
(3)
where f is the infiltration, fc is the maximum infiltration rate, f0 is the minimum infiltration rate, and k is the attenuation coefficient.

As free software, the SWMM is widely used by lots of scholars, due to its outstanding computing performance and flexible structure. Currently, the SWMM is widely applied in urban storm water simulations. Guan et al. (2015) used the SWMM to explore and assess how gradual hydrological changes occur during urban development from rural areas to a medium-density residential catchment. They found that urbanization in the residential catchment expanded the runoff contributing area, accelerated the hydrological response, and raised the peak flows. Zhu et al. (2016) proposed an approach to qualitatively and quantitatively evaluate inundation risks in urban drainage systems based on the SWMM. Xiong et al. (2019) established the SWMM of a representative urban area in Wuhan city, China, to investigate the response of drainage infrastructure to future design rainfall. The results showed that the incapability of the current drainage infrastructure in the study area was aggravated under future climate conditions.

The SWMM has been updated several times since its development in 1971. Starting with SWMM version 5.0, the LID module was added, which can flexibly generalize LIDs that are represented by a combination of vertical layers whose properties (such as thickness, void volume, hydraulic conductivity, and underdrain characteristics) are defined on a per-unit-area basis, and LIDs can be assigned within selected sub-catchments by defining the corresponding areal coverage (Mei et al. 2018; Zhu et al. 2019). Studies using the SWMM for LIDs have been extensive in different regions and countries. Li et al. (2019) used the SWMM to evaluate the comprehensive performance of LID practices for sponge city construction in Guangxi, China. The results showed that the green scenario, which contained 34.5% bioretention facilities and 46.0% sunken green spaces, had the most comprehensive performance. Cipolla et al. (2016) conducted a long-term hydrological simulation experiment on green roofs through the SWMM, and the results confirmed that green roofs function well in restoring natural water. Zeng et al. (2019) used the SWMM to discuss the role of drainage networks in LID planning and explore the characteristics of the synergistic effect. The results confirmed the existence of a synergistic effect between LID facilities and the drainage networks.

As the SWMM has been well applied in both urban storm water simulations and LID evaluations, it was selected as the simulation model in this study. The SWMM version 5.1.013 used in this study is the latest version that was released by the EPA official website on July 31, 2018.

Model generalization

In recent years, studies using a combination of the SWMM and ArcGIS tools to generalize models have achieved some successful results (Guan et al. 2015). In this study, there are four steps to generalize the SWMM model using ArcGIS tools. First, the junctions of the drainage system were generalized as nodes, and the conduits were generalized as lines in ArcGIS. Second, the Thiessen tool in ArcGIS was used to generate Tyson polygons, which were centered on each node. Third, the Tyson polygons were adjusted according to the boundary conditions, building distribution, gradient, and aspect of the study area. We took the adjusted Tyson polygons as the sub-catchments. At the same time, the outlet of each sub-catchment was designated. Fourth, data on the study area, such as the elevation of junctions, length of conduits, area, width, impervious percentage, and slope of sub-catchments were extracted by ArcGIS. Finally, the generalized model was imported into the SWMM by inp.PINS software, which can use the information contained in ArcGIS to generate input files for the SWMM. Consequently, 120 sub-catchments, 118 junctions, 120 conduits, and 3 outfalls were defined (Figure 4).

Model parameters

The parameters of the SWMM model are divided into deterministic parameters and uncertain parameters in this research (Table 1). The deterministic parameters, such as the slope, impervious percentage of each sub-catchment, elevation of each junction invert, cross-section geometry, length, and roughness of each conduit, are obtained from the documents listed in the section ‘Data collection’. For uncertain parameters, related studies (Choi & Ball 2002; Zhao et al. 2013) indicated that fc, f0, and k of the Horton model, the Manning sensitive parameters coefficient (N), and the depression storage (D) of impervious and pervious areas are sensitive parameters in the SWMM. These parameters are obtained from calibration and validation in the section ‘Calibration and validation’. The initial values of these sensitive parameters shown in Table 2 are set up before calibration and validation according to Rossman (2014), Luan et al. (2017), and Bai et al. (2019).

Table 1

The classification of SWMM parameters

Deterministic parametersUncertain parameters
Sub-catchment Area; Width; Slope; Impervious percentage N of impervious area; N of pervious area; D of impervious area; D of pervious area 
Horton fc; f0; k 
Conduit Cross section geometry; Length; Inlet offset; Outlet offset Roughness 
Junction Elevation – 
Deterministic parametersUncertain parameters
Sub-catchment Area; Width; Slope; Impervious percentage N of impervious area; N of pervious area; D of impervious area; D of pervious area 
Horton fc; f0; k 
Conduit Cross section geometry; Length; Inlet offset; Outlet offset Roughness 
Junction Elevation – 
Table 2

Best-fit model parameters of this study

Parameters
Initial valuesValidated values
N of impervious area 0.011 0.015 
N of pervious area 0.120 0.240 
D of impervious area 0.05 2.5 
D of pervious area 4.5 
Roughness of conduits 0.013 0.013 
Horton fc 10 75.6 
f0 10.5 
k 7.4 
Parameters
Initial valuesValidated values
N of impervious area 0.011 0.015 
N of pervious area 0.120 0.240 
D of impervious area 0.05 2.5 
D of pervious area 4.5 
Roughness of conduits 0.013 0.013 
Horton fc 10 75.6 
f0 10.5 
k 7.4 

Calibration and validation

Because the observed discharge data are absent in the CDLC, the MSDs of the typical precipitation processes (the ‘7.19’ and ‘5.22’ storms) were selected as the key factors for model calibration and validation (Luan et al. 2017). The relative error (RE) of the simulated MSD (hereafter, MSDs) and the IS MSD (hereafter, MSDI) were used as the check factors to calibrate and validate the generalized model. The RE solution formula is shown in Equation (4):
formula
(4)

The values of MSDs were obtained from the simulation results of the ‘7.19’ and ‘5.22’ precipitation events, which were input into the generalized model. The MSDI values of the ‘7.19’ and ‘5.22’ precipitation processes were obtained from IS.

The MSDI, MSDS, and RE data of the ‘7.19’ storm were chosen for the model calibration. After several calibration interactions, the optimal parameter values were finally obtained (Table 2) according to the minimum RE (Table 3) in each IS position and its simulated runoff process is shown in Figure 5(a). Then, we validated the simulation with the MSDI, MSDS, and RE data of the ‘5.22’ storm. Its simulated runoff process and RE are shown in Figure 5(b) and Table 3, respectively. From Figure 5, we can see that the simulated runoff processes of ‘7.19’ and ‘5.22’ storms are closely responsive to the rainfall process. Table 3 shows that the simulations demonstrate an RE that ranges from −9.4% to 10.7% (‘7.19’ and ‘5.22’ rainfall events, respectively). The results indicated that the model simulation results were consistent with the investigation data, according to the national standard (Supervision 2008; Luan et al. 2017). Thus, the calibration and validation results verified that the developed model performs well in modeling the hydrological responses in the study area, and that it has potential for use in further scenario analyses. The best-fit model parameters are shown in Table 2.

Table 3

Comparison between the MRDI and MRDS values of the ‘7.19’ and ‘5.22’ storms

Position‘7.19’ storm
Position‘5.22’ storm
MSDI (cm)MSDS (cm)RE (%)MSDI (cm)MSDS (cm)RE (%)
1# 10 11.01 10.1 1# 8.5 7.70 − 9.4 
2# 13 13.11 0.9 2# 7.0 7.48 6.9 
3# 9.53 5.9 3# 8.0 7.98 − 0.3 
4# 16 17.04 6.5 4# 6.0 5.59 − 6.8 
5# 10 10.73 7.3 5# 4.0 4.43 10.7 
Position‘7.19’ storm
Position‘5.22’ storm
MSDI (cm)MSDS (cm)RE (%)MSDI (cm)MSDS (cm)RE (%)
1# 10 11.01 10.1 1# 8.5 7.70 − 9.4 
2# 13 13.11 0.9 2# 7.0 7.48 6.9 
3# 9.53 5.9 3# 8.0 7.98 − 0.3 
4# 16 17.04 6.5 4# 6.0 5.59 − 6.8 
5# 10 10.73 7.3 5# 4.0 4.43 10.7 
Figure 5

Simulated runoff processes of the (a) ‘7.19’ and (b) ‘5.22’ storms.

Figure 5

Simulated runoff processes of the (a) ‘7.19’ and (b) ‘5.22’ storms.

Simulation scenarios

LID scenario

In this study, the hypothetical numbers, types, and locations of LIDs were determined primarily based on the land use characteristics (Figure 2) of the study area. Furthermore, the ‘Technical Guide for Sponge Cities-Construction of Low-Impact Development’ (hereafter, TGSC) was referenced to determine the criteria for LID design (MHURD 2014).

Figure 2 shows that the land use characteristics of the study area are mainly divided into four categories: residential land, commercial land, industrial land, and greening land, which comprise 30, 12.5, 44.5, and 13% of the total area, respectively. The types, purposes and characteristics of these land uses are shown in Figure 6. Residential land is a living area, mainly including old housing, newly built housing, and indemnificatory housing, with a high building density and a high proportion of impervious area. Commercial land is centered on public services and civic activities, including cultural entertainment, commercial and medical facilities, with a high building density, more hardened squares and pavements, and a high proportion of impervious areas. Industrial land is a centralized area of industrial and logistic warehouses with a large workshop area, high building density, and high proportion of impervious area. Greening land is mainly composed of the greenbelts of parks, with fewer buildings, more permeable surfaces, and a larger proportion of pervious area.

Figure 6

The types, purposes, and characteristics of land use in the study area.

Figure 6

The types, purposes, and characteristics of land use in the study area.

Based on a comprehensive consideration of land use characteristics (Figure 1), topographic conditions (Figure 2), and the area that can be modified, LIDs were set in the SWMM based on the purpose, applicability, and arrangement principles of different LIDs in the TGSC. The specific methods are as follows: (1) In residential land and industrial land, where the terrain is gentle and the slope is low, green roofs and vegetated swales were implemented. Green roofs are used to reduce building roof runoff, and residual roof runoff is transferred to the surrounding pervious areas through vegetated swales. (2) In commercial land in the middle area and east area, where the terrain is steep, due to the large number of squares and roads, permeable pavements were implemented to increase the proportion of pervious areas, improve the infiltration capacity of the underlying surface, and reduce the urban waterlogging pressure on the downstream area. (3) The greening land is designed as a sink (i.e., concave greenbelts) to increase the storage capacity of rainwater.

After setting the LIDs described above, the designed parameters of the LIDs were arranged in the model based on the TGSC and previous studies (Rossman 2014; Palla & Gnecco 2015; Shen et al. 2018). The setting proportion and main parameters of the LIDs are shown in Table 4. An illustration of the distribution of LIDs is shown in Figure 7.

Table 4

The designed proportion and main parameters of LIDs

Proportions (%)
Parameters
Types of LIDsWest areaMiddle areaEast areaLayersSurfaceSoilPavementStorageDrainage
Green roofs 14.72 7.52 16.56 Thickness (mm) 100 150 – – 30 
Roughness 0.24 – – – 0.1 
Conductivity (mm/h) –- 0.5 – – – 
Vegetated swales 7.29 3.76 8.52 Thickness (mm) 200 – – – – 
Roughness 0.24 – – – – 
Concave greenbelts – 21.28 8.96 Thickness (mm) 200 400 – 100 – 
Roughness 0.24 – – – – 
Conductivity (mm/h) – 0.5 – – – 
Void ratio – – – 0.75 – 
Flow coefficient – – – – 
Permeable pavements – 1.02 1.78 Thickness (mm) 200 – 100 300 – 
Roughness 0.24 – – –  
Conductivity (mm/h) – 0.5 – – – 
Void ratio – – 0.15 0.52 – 
Flow coefficient – – – – 
Proportions (%)
Parameters
Types of LIDsWest areaMiddle areaEast areaLayersSurfaceSoilPavementStorageDrainage
Green roofs 14.72 7.52 16.56 Thickness (mm) 100 150 – – 30 
Roughness 0.24 – – – 0.1 
Conductivity (mm/h) –- 0.5 – – – 
Vegetated swales 7.29 3.76 8.52 Thickness (mm) 200 – – – – 
Roughness 0.24 – – – – 
Concave greenbelts – 21.28 8.96 Thickness (mm) 200 400 – 100 – 
Roughness 0.24 – – – – 
Conductivity (mm/h) – 0.5 – – – 
Void ratio – – – 0.75 – 
Flow coefficient – – – – 
Permeable pavements – 1.02 1.78 Thickness (mm) 200 – 100 300 – 
Roughness 0.24 – – –  
Conductivity (mm/h) – 0.5 – – – 
Void ratio – – 0.15 0.52 – 
Flow coefficient – – – – 
Figure 7

Illustration of the distribution of LID facilities.

Figure 7

Illustration of the distribution of LID facilities.

Based on the methods for setting the LIDs described above, the LIDs were proposed. The status quo scenario was defined as the baseline (BL), against which the effectiveness of the LID scenario in terms of urban waterlogging mitigation was quantified.

Designed storm scenarios

A design rainstorm is the basis of urban rainwater pipe network planning, which serves the needs of local flooding and waterlogging control and drainage projects. There are many design rainfall models, such as the Chicago rainfall model (CRM), uniform rainfall model, Huff rainfall model, and triangle rainfall model (Jun et al. 2018). The CRM, which was proposed by Keifer & Chu (1957), is an uneven rainfall model based on the intensity–duration–frequency relationship. It is widely used in urban water resource engineering. The CRM is based on a rainfall intensity formula to design typical precipitation processes. In this project, the rainfall intensity formula is summarized by Equation (5):
formula
(5)
where i is the rainfall intensity (mm/s), t is the rainfall duration (h), and P is the return period (a). In the CRM, the rainfall return period P, rainfall duration t, and rainfall peak coefficient r (0 < r < 1) are the main factors determining the precipitation processes. Among them, the rainfall peak coefficient r is defined as the ratio of the time before the peak intensity to the total duration. The variable r describes the location of peak rainfall intensity: the smaller r is, the closer the peak intensity is to the rainfall starting time.

To explore the controlling runoff performance of LIDs under different storm conditions, three groups (Figure 8) were considered in this study. In Group 1 (Figure 8(a)), the storm events have different return periods (2a, 5a, 10a, 20a, and 50a), and the corresponding total rainfall amounts range from 52.27 to 106.07 mm. These storms all have the same rainfall duration (2 h) and location of peak rainfall intensity (r = 0.4). In Group 2 (Figure 8(b)), the storm events have different rainfall durations (2 h, 4 h, 6 h, and 8 h). They have the same return period (10a) and location of peak intensity (r = 0.4). In Group 3 (Figure 8(c)), the storm events have different rainfall peak coefficients r (0.2, 0.4, 0.6, and 0.8), and they have the same return period (10a) and rainfall duration (2 h). The aims of Group 1, Group 2, and Group 3 are to investigate how the performances of LIDs, in terms of urban waterlogging, are affected by rainfall intensity, rainfall duration, and the location of peak intensity, respectively, in semi-mountainous regions.

Figure 8

Design storms for scenario analysis.

Figure 8

Design storms for scenario analysis.

Hydrological indices for urban waterlogging mitigation assessments

For evaluating the effectiveness of the LID scenario for urban waterlogging mitigation under different storm conditions, three hydrological performance indices were selected for this research in accordance with previous studies (Mei et al. 2018; Hu et al. 2019), i.e., the reduction ratios of surface runoff (RW), the reduction ratios of peak flow (RP), and the time delay of peak time (DT), which were calculated based on comparisons between the hydrological results of the LID and BL scenarios. Here, surface runoff refers to the total amount of the runoff system, peak flow refers to the highest flow rate of the runoff system, and peak time refers to the moment when the highest flow rate of the runoff system occurs. The RW, RP, and DT indices, which were calculated through Equations (6)–(8), respectively, reflect the effects of LIDs on the rainfall-runoff process. The three selected indices were considered to constitute a reasonably complete combination for the evaluation of the runoff control effectiveness of LIDs. The values of surface runoff, peak flow, and peak time were output by the SWMM in the summary files of the simulations.
formula
(6)
formula
(7)
formula
(8)
where WB and WL are the surface runoffs of the BL and LID scenarios, respectively (m3), PB and PL are the peak flows of the BL and LID scenarios, respectively (m3/s), and TB and TL are the peak time of the BL and LID scenarios, respectively (min).

RESULTS ANALYSIS AND DISCUSSION

The runoff response process of CDLC was simulated in the BL and LID scenarios with different rainfall characteristics through the calibrated SWMM. The flow process of outlets O1, O2, and O3, which controls the west, middle, and east areas, respectively, was compared and analyzed as follows. Furthermore, the effectiveness of regional LIDs with different rainfall characteristics was accurately assessed through surface runoff reduction, peak flow diminution, and delay duration of peak flow time.

Analysis and comparison of the flow process of each outlet

The flow process under different rainfall return periods

The simulated flow processes of O1, O2, and O3 under different return periods in the BL and LID scenarios are shown in Figure 9.

Figure 9

The flow processes of each outlet under different rainfall return periods.

Figure 9

The flow processes of each outlet under different rainfall return periods.

It is indicated in Figure 9 that all the flow processes are sharp and steep in the BL scenario under different return periods (from 2a to 20a). All of the flow processes in the LID scenario are shorter and flatter than those in the BL scenario under different return periods. As the return periods increase, all of the flow processes in both BL and LID scenarios become steeper.

It should be noted that the flow process of O1 in LID scenario under 5a return period basically coincides with that in the BL scenario under 2a return period, as shown in Figure 9(a). In particular, the receding flow process of O1 after 3:45 in the LID scenario under 5a return period is below that in the BL scenario under 2a return period. From Figure 9(b), the flow process of O2 in the LID scenario under 20a return period basically coincides with that in the BL scenario under 5a return period, and the peak flow in the LID scenario under 20a return period is smaller than that in the BL scenario under 5a return period. The flow process of O2 in LID scenario under 5a return period basically coincides with that in the BL scenario under 2a return period, and the peak flow and receding flow process in the LID scenario under 5a return period is below that in the BL scenario under 2a return period. Similarly, in Figure 9(c), the flow process of O3 in the LID scenario under 10a return period basically coincides with that in the BL scenario under 2a return period, and both of the peak flows are approximately equal.

The flow process under different rainfall durations

The simulation flow processes of O1, O2, and O3 under different rainfall durations in the BL and LID scenarios are shown in Figure 10.

Figure 10

The flow processes of each outlet under different rainfall durations.

Figure 10

The flow processes of each outlet under different rainfall durations.

Figure 10 shows that all of the flow processes of each outlet are sharp and steep in the BL scenario under different rainfall durations (from 2 to 8 h). All of the flow processes of each outlet in the LID scenario are gentler than those in the BL scenario under all rainfall durations. As the rainfall duration increases, all of the peaks of the flow processes for each outlet in both the BL and LID scenarios increase.

As the rainfall duration increases, it is worth noting that the flow processes of O1 and O3 become smoother in the LID scenario, especially the peak of the flow process, as shown in Figure 10(a) and 10(c). However, Figure 10(b) shows that the flow processes of O2 have no such effect. The peak of the flow process of O2 in the LID scenario under the 8 h rainfall duration is steeper than that under the 6 and 4 h rainfall durations, and the peak of the flow process of O2 in the LID scenario under the 6 h rainfall duration is steeper than that under the 4 h rainfall duration, which may be related to the slope. From Figure 10, the middle area has a greater slope than that in the west area and east area. This shows that the regulation and storage effects of LIDs are weak in areas with large slopes.

The flow process of different rainfall peak coefficients

The simulation flow processes of O1, O2, and O3 under different rainfall peak coefficients in the BL and LID scenarios are shown in Figure 11.

Figure 11

The flow processes of each outlet under different rainfall peak coefficients.

Figure 11

The flow processes of each outlet under different rainfall peak coefficients.

Similarly, Figure 11 shows that all of the flow processes of each outlet are sharp and steep in the BL scenario under different rainfall peak coefficients (from 0.2 to 0.8). All of the flow processes of each outlet in the LID scenario are gentler than those in the BL scenario under all rainfall peak coefficients.

Figure 11(a) shows that the flow process lines of O1 are relatively close both in the BL and LID scenarios under different rainfall peak coefficients, which is also true for O2 and O3, as shown in Figure 11(b) and 11(c), respectively. When the rainfall peak coefficient increases from 0.2 to 0.8, the rising time of the flow process at O1, O2, and O3 slightly lags behind both in the BL and LID scenarios, and the peaks in the LID scenario are extremely close, especially the flow process of O3. Furthermore, the degrading process of the flow process lines at each outlet almost coincides in both the BL and LID scenarios as the rainfall peak coefficient increases from 0.2 to 0.8.

Rainfall characteristics impact on the urban waterlogging reduction

Impact of rainfall return period

Table 5 shows the simulated surface runoff, peak flow, and peak time in the BL and LID scenarios under different designed rainfall return periods. Figure 12 shows the reduction trends of surface runoff and peak flow with rainfall return period.

Table 5

Surface runoff, peak flow, and peak time in the BL and LID scenarios under different rainfall return periods

Surface runoff
Peak flow
Delay of peak time
PBL/m3LID/m3RW /%BL/m3·s−1LID/ m3·s−1RP/%BL/hh:mmLID/hh:mmDT/min
2a 286,610 153,510 46.44 13.04 7.73 40.72 2:01 2:01 
5a 392,080 212,880 45.71 19.49 11.62 40.38 2:01 2:01 
10a 474,420 261,750 44.83 24.85 14.86 40.20 2:01 2:01 
20a 558,750 319,480 42.82 30.56 18.41 39.76 2:01 2:01 
Surface runoff
Peak flow
Delay of peak time
PBL/m3LID/m3RW /%BL/m3·s−1LID/ m3·s−1RP/%BL/hh:mmLID/hh:mmDT/min
2a 286,610 153,510 46.44 13.04 7.73 40.72 2:01 2:01 
5a 392,080 212,880 45.71 19.49 11.62 40.38 2:01 2:01 
10a 474,420 261,750 44.83 24.85 14.86 40.20 2:01 2:01 
20a 558,750 319,480 42.82 30.56 18.41 39.76 2:01 2:01 
Figure 12

The trends of RW and RP with rainfall return period.

Figure 12

The trends of RW and RP with rainfall return period.

From Table 5, it is found that the reduction effect of LIDs is obvious under the design rainfall of Group 1 in this study area. The reductions in surface runoff and peak flow can reach 46.44% and 40.72%, respectively.

As shown in Figure 12, for the same rainfall duration and peak coefficient, the reduction ratios of surface runoff and peak flow decreased with increasing rainfall return period. As the rainfall return period increases from 2a to 20a, the reduction in surface runoff decreases from 46.44% to 42.82%, and the reduction in peak flow decreases from 40.72% to 39.76%. However, interestingly, with the increase in the rainfall return period, the reduction ratio of surface runoff decreased by 3.62%, while the reduction ratio of peak flow was almost unchanged and only decreased by 0.96%. It can be seen that the ability of LIDs to control surface runoff is more sensitive to the rainfall duration of the rainfall return period in semi-mountainous regions compared to the peak flow, which is contrary to that in plain areas (Xiang et al. 2017; Hu et al. 2019). The reason may be related to the slope. In semi-mountainous regions, the slope is larger than in plain regions. When the rainfall amount is small, surface runoff can be generated. At this time, the storage capacity of LIDs is not saturated. LIDs also retain a portion of the water storage capacity to cope with the peak flow.

It is indicated in Table 5 that the peak time is not delayed under all of Group 1 rainfall, and the peak time occurs at 2:01. The main reason is that the semi-mountainous region has a large slope and steep terrain, and rainwater is more likely to be lost to form surface runoff.

Impact of rainfall duration

Table 6 shows the simulated surface runoff, peak flow and peak time in the BL and LID scenarios under different designed rainfall durations. Figure 13 shows the reduction trends of surface runoff and peak flow with rainfall duration.

Table 6

Surface runoff, peak flow and peak time in the BL and LID scenarios under different rainfall durations

Surface runoff
Peak flow
Delay of peak time
tBL/m3LID/m3RW /%BL/m3·s−1LID/m3·s−1RP/%BL/hh:mmLID/hh:mmDT/min
2 h 474,420 261,750 44.83 24.85 14.86 40.20 2:01 2:01 
4 h 626,570 369,800 40.98 28.44 17.01 40.19 2:40 2:46 
6 h 726,720 449,930 38.09 30.83 18.47 40.09 3:24 3:32 
8 h 801,130 512,490 36.03 32.61 19.58 39.96 4:06 4:16 10 
Surface runoff
Peak flow
Delay of peak time
tBL/m3LID/m3RW /%BL/m3·s−1LID/m3·s−1RP/%BL/hh:mmLID/hh:mmDT/min
2 h 474,420 261,750 44.83 24.85 14.86 40.20 2:01 2:01 
4 h 626,570 369,800 40.98 28.44 17.01 40.19 2:40 2:46 
6 h 726,720 449,930 38.09 30.83 18.47 40.09 3:24 3:32 
8 h 801,130 512,490 36.03 32.61 19.58 39.96 4:06 4:16 10 
Figure 13

The trends of RW and RP with rainfall duration.

Figure 13

The trends of RW and RP with rainfall duration.

Table 6 shows that the reduction effect of LIDs is obvious under the design rainfall of Group 2 in this study area. The reduction in surface runoff and peak flow can reach 44.83% and 40.20%, respectively.

As shown in Figure 13, under the same rainfall return period and peak coefficient, the reduction ratio of surface runoff decreases with the increase in rainfall duration. As the rainfall duration extends from 2 to 8 h, the reduction in surface runoff decreases from 44.83% to 36.03%. This conclusion is in contrast to that in the plain area (Xiang et al. 2017; Hu et al. 2019). In these previous conclusions, the surface runoff increases with the increase in rainfall duration. In the plain region, as the rainfall duration increases, the same time distribution of rainfall is stable, so more rainfall can infiltrate into the LIDs, instead of quickly overflowing. However, in mountainous or semi-mountainous areas, even if the rainfall distribution is stable, it will flow away quickly due to the large slope of the area.

From Table 6 and Figure 13, it can be seen that the influence of rainfall duration on the peak flow reduction and peak time delay of LIDs is small. As rainfall duration increases from 2 to 8 h, the reduction rate of the peak flow decreases, with a small range of only 0.24%. The longest peak time delay is only 10 min when the rainfall duration is 8 h.

Impact of rainfall peak coefficient

Table 7 shows the simulated surface runoff, peak flow, and peak time in the BL and LID scenarios under different design rainfall peak coefficients. Figure 14 shows the reduction trends of surface runoff and peak flow with the rainfall peak coefficient.

Table 7

Surface runoff, peak flow, and peak time in the BL and LID scenarios under different rainfall peak coefficients

Surface runoff
Peak flow
Delay of peak time
rBL/m3LID/m3RW/%BL/m3·s−1LID/m3·s−1RP/%BL/hh:mmLID/hh:mmDT/min
0.2 475,970 262,790 44.79 23.68 14.15 40.25 2:01 2:01 
0.4 474,420 261,750 44.83 24.85 14.86 40.20 2:01 2:01 
0.6 473,120 260,670 44.90 26.42 15.8 40.20 1:57 1:57 
0.8 472,160 259,690 44.80 28.62 17.07 40.36 1:56 1:56 
Surface runoff
Peak flow
Delay of peak time
rBL/m3LID/m3RW/%BL/m3·s−1LID/m3·s−1RP/%BL/hh:mmLID/hh:mmDT/min
0.2 475,970 262,790 44.79 23.68 14.15 40.25 2:01 2:01 
0.4 474,420 261,750 44.83 24.85 14.86 40.20 2:01 2:01 
0.6 473,120 260,670 44.90 26.42 15.8 40.20 1:57 1:57 
0.8 472,160 259,690 44.80 28.62 17.07 40.36 1:56 1:56 
Figure 14

The trends of RW and RP with rainfall duration.

Figure 14

The trends of RW and RP with rainfall duration.

From Table 7, it is found that the reduction effect of LIDs is obvious under the design rainfall of Group 3 in this study area. The reductions in surface runoff and peak flow can reach 44.79% and 40.36%, respectively.

As shown in Figure 14, under the same rainfall return period and duration, there is no significant change in the reduction ratio of surface runoff with an increasing rainfall peak coefficient. As the rainfall peak coefficient increases from 0.2 to 0.8, the reduction in surface runoff only changes by 0.11%. Similarly, there is no significant change in the reduction ratio of peak flow with increasing rainfall peak coefficient, only 0.16%, which is contrary to that in plain areas (Qin et al. 2013; Xiang et al. 2017). In these previous conclusions, the reduction ratio of the peak flow decreased significantly with increasing rainfall peak coefficient. In the plain regions, when the rainfall reached the later stage, the soil infiltration rate was smaller, and the storage capacity of LIDs was more saturated. Therefore, LIDs have less ability to reduce peak flow when peak rainfall occurs. However, in mountainous or semi-mountainous areas, the slope is larger than that in the plain regions. When the rainfall is in the early stage, surface runoff can be generated. At this time, the storage capacity of LIDs is not saturated. When the rainfall reaches the later stage, LIDs also retain a portion of the water storage capacity to cope with the peak flow. Therefore, the reduction ratio of peak flow does not significantly change with the delay of peak flow time.

It is indicated in Table 7 that the peak time does not delay under Group 3 rainfall, and it occurs at 2:01, 2:01, 1:56, and 1:57.

This study has some shortcomings, shown as follows, which will be considered in future research. Due to the absence of the observed pipeline discharge, investigation is applied for model calibration and validation. In future, pipeline discharge could be measured and utilized in order to improve the model accuracy. The purpose of this study is to assess the simulated LID effectiveness; however, in the actual engineering application, LID location in the residential area and the related investment should be considered. In addition, more LID varieties also should be considered according to the local condition, and the LID effectiveness in other semi-mountainous or mountainous areas should be assessed and researched. Furthermore, the response characteristics of the catchment changed by LID implementation were not considered in this paper due to the complex computation and space limitation, besides the main content of this paper's emphasis on the evaluation of the runoff response to the different storm characteristics in semi-mountainous regions. This content will be supplemented in future research.

CONCLUSIONS

This study analyzed the LID effectiveness of semi-mountainous regions on urban waterlogging reduction in various designed storm scenarios, and the corresponding results indicated that all of the LIDs in different storm scenarios were effective. The surface runoff and peak flow in different LID scenarios decreased by 36.03–46.44% and 39.76–40.72%, respectively, comparing the flow processes in the BL scenario. With the influence of slope, the LID effectiveness in semi-mountainous regions on urban waterlogging reduction under various designed storm scenarios is different from that in plain regions. The details are as follows.

As the rainfall return period increases from two years to twenty years, the reduction rate of surface runoff decreases by 3.76%. As the rainfall duration increases from 2 to 8 h, the reduction ratio of surface runoff decreases by 8.8%, but the reduction rate of surface runoff is little, only 0.11%, with the delay of the peak flow time.

However, the LID effectiveness on controlling peak flow in semi-mountainous regions is not remarkable: the reduction rate decreases by 0.96% as the rainfall return period increases from two years to twenty years; decreases by only 0.24% as the rainfall duration increases from 2 to 8 h; and decreases by only 0.16% as the relative peak location of flow process changed from 0.2 to 0.8. Furthermore, the regional LID effectiveness of delaying peak time is very limited: the longest delay time is only 10 min as the rainfall duration is 8 h.

This research can provide technical support for LID implementation in semi-mountainous regions. Decision-makers can make reasonable designs and plans according to this study to reduce the flooding and waterlogging disasters in mountainous or semi-mountainous cities.

ACKNOWLEDGEMENT

This study was supported by the Chinese National Natural Science Foundation (No. 51739011), the Key Research and Development Project of Hebei Province (20375401D) and the Research Fund of the State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin (SKL2020ZY03). Authors' contributions are as follows: methodology, D.W., Q.L., J.L. and H.W.; software and validation, D.W. and X.F.; investigation, resources, and data curation, Q.L. and D.W.; writing – original draft preparation, D.W.; writing – review and editing, all authors; visualization, D.W. and X.F.; supervision, Q.L.; project administration and overall guidance, J.L. and H.W.; funding acquisition, J.L. All authors have read and agreed to the published version of the manuscript. The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT

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

REFERENCES

REFERENCES
Ahiablame
L. M.
,
Engel
B. A.
&
Chaubey
I.
2012
Effectiveness of low impact development practices: literature review and suggestions for future research
.
Water Air and Soil Pollution
223
(
7
),
4253
4273
.
doi:10.1007/s11270-012-1189-2
.
Alfredo
K.
,
Montalto
F.
&
Goldstein
A.
2010
Observed and modeled performances of prototype green roof test plots subjected to simulated low- and high-intensity precipitations in a laboratory experiment
.
Journal of Hydrologic Engineering
15
(
6SI
),
444
457
.
doi:10.1061/(ASCE)HE.1943-5584.0000135
.
Bai
Y.
,
Zhao
N.
,
Zhang
R.
&
Zeng
X.
2019
Storm water management of low impact development in urban areas based on SWMM
.
Water
11
(
331
).
doi:10.3390/w11010033
.
Bedan
E. S.
&
Clausen
J. C.
2009
Stormwater runoff quality and quantity from traditional and low impact development watersheds
.
Journal of the American Water Resources Association
45
(
4
),
998
1008
.
doi:10.1111/j.1752-1688.2009.00342.x
.
Chen
Y.
,
Samuelson
H. W.
&
Tong
Z.
2016
Integrated design workflow and a new tool for urban rainwater management
.
Journal of Environmental Management
180
,
45
51
.
doi:10.1016/j.jenvman.2016.04.059
.
Choi
K.
&
Ball
J. E.
2002
Parameter estimation for urban runoff modelling
.
Urban Water
4
(
1
),
31
41
.
doi:10.1016/S1462-0758(01)00072-3
.
Cipolla
S. S.
,
Maglionico
M.
&
Stojkov
I.
2016
A long-term hydrological modelling of an extensive green roof by means of SWMM
.
Ecological Engineering
95
,
876
887
.
doi:10.1016/j.ecoleng.2016.07.009
.
Dietz
M. E.
&
Clausen
J. C.
2008
Stormwater runoff and export changes with development in a traditional and low impact subdivision
.
Journal of Environmental Management
87
(
4
),
560
566
.
doi:10.1016/j.jenvman.2007.03.026
.
Eckart
K.
,
McPhee
Z.
&
Bolisetti
T.
2017
Performance and implementation of low impact development-a review
.
Science of the Total Environment
607–608
,
413
432
.
doi:10.1016/j.scitotenv.2017.06.254
.
Fletcher
T. D.
,
Shuster
W.
,
Hunt
W. F.
,
Ashley
R.
,
Butler
D.
,
Arthur
S.
,
Trowsdale
S.
,
Barraud
S.
,
Annette
S.
,
Mikkelsen
P. S.
,
Rivard
G.
,
Mathias
U.
,
Danielle
D.
&
Maria
V.
2015
SUDS, LID, BMPs, WSUD and more-The evolution and application of terminology surrounding urban drainage
.
Urban Water Journal
12
(
7
),
525
542
.
doi:10.1080/1573062X.2014.916314
.
Guan
M.
,
Sillanpaa
N.
&
Koivusalo
H.
2015
Modelling and assessment of hydrological changes in a developing urban catchment
.
Hydrological Processes
29
(
13
),
2880
2894
.
doi:10.1002/hyp.10410
.
Gunn
R.
,
Martin
A.
,
Engel
B.
&
Ahiablame
L.
2012
Development of two indices for determining hydrologic implications of land use changes in urban areas
.
Urban Water Journal
9
(
4
),
239
248
.
doi:10.1080/1573062X.2012.660957
.
Hood
M. J.
,
Clausen
J. C.
&
Warner
G. S.
2007
Comparison of stormwater lag times for low impact and traditional residential development
.
Journal of the American Water Resources Association
43
(
4
),
1036
1046
.
doi:10.1111/j.1752-1688.2007.00085.x
.
Hu
M.
,
Sayama
T.
,
Zhang
X.
,
Tanaka
K.
,
Takara
K.
&
Yang
H.
2017
Evaluation of low impact development approach for mitigating flood inundation at a watershed scale in China
.
Journal of Environmental Management
193
,
430
438
.
doi:10.1016/j.jenvman.2017.02.020
.
Hu
M.
,
Zhang
X.
,
Li
Y.
,
Yang
H.
&
Tanaka
K.
2019
Flood mitigation performance of low impact development technologies under different storms for retrofitting an urbanized area
.
Journal of Cleaner Production
222
,
373
380
.
doi:10.1016/j.jclepro.2019.03.044
.
Jacobson
C. R.
2011
Identification and quantification of the hydrological impacts of imperviousness in urban catchments: a review
.
Journal of Environmental Management
92
(
6
),
1438
1448
.
doi:10.1016/j.jenvman.2011.01.018
.
Jun
L.
,
Wu
S.
,
Zhao
X.
,
Yi
W.
,
Zhang
S.
,
Li
M.
&
Li
J.
2018
Analysis on the effect of rain type selection on LID measures
.
Water & Wastewater Engineering
54
(
5
),
21
27
.
(in Chinese)
.
doi:10.13789/j.cnki.wwe1964.2018.0097
.
Keifer
C. J.
&
Chu
H. H.
1957
Synthetic storm pattern for drainage design
.
Journal of the Hydraulics Division
83
(
4
),
1
25
.
Kong
F.
,
Ban
Y.
,
Yin
H.
,
James
P.
&
Dronova
I.
2017
Modeling stormwater management at the city district level in response to changes in land use and low impact development
.
Environmental Modelling & Software
95
,
132
142
.
doi:10.1016/j.envsoft.2017.06.021
.
Li
J.
,
Zhao
R.
,
Li
Y.
&
Chen
L.
2018
Modeling the effects of parameter optimization on three bioretention tanks using the HYDRUS-1D model
.
Journal of Environmental Management
217
,
38
46
.
doi:10.1016/j.jenvman.2018.03.078
.
Li
Q.
,
Wang
F.
,
Yu
Y.
,
Huang
Z.
,
Li
M.
&
Guan
Y.
2019
Comprehensive performance evaluation of LID practices for the sponge city construction: a case study in Guangxi, China
.
Journal of Environmental Management
231
,
10
20
.
doi:10.1016/j.jenvman.2018.10.024
.
Liu
Y.
,
Bralts
V. F.
&
Engel
B. A.
2015
Evaluating the effectiveness of management practices on hydrology and water quality at watershed scale with a rainfall-runoff model
.
Science of the Total Environment
511
,
298
308
.
doi:10.1016/j.scitotenv.2014.12.077
.
Luan
Q.
,
Fu
X.
,
Song
C.
,
Wang
H.
,
Liu
J.
&
Wang
Y.
2017
Runoff effect evaluation of LID through SWMM in typical mountainous, low-lying urban areas: a case study in China
.
Water
9
(
4396
).
doi:10.3390/w9060439
.
Maniquiz-Redillas
M. C.
&
Kim
L.
2016
Evaluation of the capability of low-impact development practices for the removal of heavy metal from urban stormwater runoff
.
Environmental Technology
37
(
18
),
2265
2272
.
doi:10.1080/09593330.2016.1147610
.
Mei
C.
,
Liu
J.
,
Wang
H.
,
Yang
Z.
,
Ding
X.
&
Shao
W.
2018
Integrated assessments of green infrastructure for flood mitigation to support robust decision-making for sponge city construction in an urbanized watershed
.
Science of the Total Environment
639
,
1394
1407
.
doi:10.1016/j.scitotenv.2018.05.199
.
MHURD
2014
Preliminary Technical Guidance for Sponge City Construction-Low Impact Development Rainwater System Construction
.
MHURD
,
Beijing
,
China
.
Palla
A.
&
Gnecco
I.
2015
Hydrologic modeling of low impact development systems at the urban catchment scale
.
Journal of Hydrology
528
,
361
368
.
doi:10.1016/j.jhydrol.2015.06.050
.
Qin
H.
,
Li
Z.
&
Fu
G.
2013
The effects of low impact development on urban flooding under different rainfall characteristics
.
Journal of Environmental Management
129
,
577
585
.
doi:10.1016/j.jenvman.2013.08.026
.
Rodríguez-Rojas
M. I.
,
Huertas-Fernández
F.
,
Moreno
B.
,
Martínez
G.
&
Grindlay
A. L.
2018
A study of the application of permeable pavements as a sustainable technique for the mitigation of soil sealing in cities: a case study in the south of Spain
.
Journal of Environmental Management
205
,
151
162
.
doi:10.1016/j.jenvman.2017.09.075
.
Rossman
L. A.
2014
Storm Water Management Model User's Manual Version 5.1
.
USEPA
,
Washington, DC
,
USA
.
Shen
H.
,
Zhang
S.
&
Xu
Z.
2018
LID facility allocation and rainfall-runoff monitoring for the Future Science Technology Park in Beijing
.
Journal of Hydraulic Engineering
49
(
8
),
937
944
.
doi:10.13243/j.cnki.slxb.20180269
.
Soulis
K. X.
,
Valiantzas
J. D.
,
Ntoulas
N.
,
Kargas
G.
&
Nektarios
P. A.
2017
Simulation of green roof runoff under different substrate depths and vegetation covers by coupling a simple conceptual and a physically based hydrological model
.
Journal of Environmental Management
200
,
434
445
.
doi:10.1016/j.jenvman.2017.06.012
.
Sun
Y.
,
Wei
X.
&
Pomeroy
C. A.
2011
Review of current research and future directions of low impact development practices for storm water
.
Advances in Water Science
22
(
2
),
287
293
.
Supervision
G. A. O. Q.
2008
Inspection and quarantine of the People's Republic of China, standardization administration of the People's Republic of China
. In
Standard for Hydrological Information and Hydrological Forecasting (GB/T 22482-2008)
.
The Ministry of Water Resources of the People's Republic of China
,
Beijing
,
China
.
Tan
S. B. K.
,
Chua
L. H. C.
,
Shuy
E. B.
,
Lo
E. Y.
&
Lim
L. W.
2008
Performances of rainfall-runoff models calibrated over single and continuous storm flow events
.
Journal of Hydrologic Engineering
13
(
7
),
597
607
.
doi:10.1061/(ASCE)1084-0699(2008)13:7(597)
.
Tang
Z.
,
Engel
B. A.
,
Pijanowski
B. C.
&
Lim
K. J.
2005
Forecasting land use change and its environmental impact at a watershed scale
.
Journal of Environmental Management
76
(
1
),
35
45
.
doi:10.1016/j.jenvman.2005.01.006
.
Traver
R. G.
&
Ebrahimian
A.
2017
Dynamic design of green stormwater infrastructure
.
Frontiers of Environmental Science & Engineering
11
(
4
),
15
.
doi:10.1007/s11783-017-0973-z
.
Wang
H.
,
Mei
C.
&
Liu
J.
2017
Systematic construction pattern of the sponge city
.
Journal of Hydraulic Engineering
48
(
9
),
1009
1014 and 1022
.
doi:10.13243/j.cnki.slxb.20170308
.
Wang
D.
,
Zhang
X.
,
Luan
Q.
,
Zhang
K.
,
Fu
X.
&
Liu
J.
2019
Evaluation on application of low impact development technology in hilly zone of semi-mountainous region
.
Water Resources and Hydropower Engineering
50
(
5
),
115
123
.
doi:10.13928/j.cnki.wrahe.2019.05.014
.
Xiang
C.
,
Liu
J.
,
Shao
W.
,
Mei
C.
,
Xia
L.
&
Wang
H.
2017
Effectiveness of systematically-installed sponge facilities under different rain patterns
.
Water Resources and Hydropower Engineering
48
(
6
),
7
12
.
(in Chinese) doi:10.13928/j.cnki.wrahe.2017.06.002
.
Xiong
L.
,
Yan
L.
,
Du
T.
,
Yan
P.
,
Li
L.
&
Xu
W.
2019
Impacts of climate change on urban extreme rainfall and drainage infrastructure performance: a case study in Wuhan City, China
.
Irrigation and Drainage
68
(
2
),
152
164
.
doi:10.1002/ird.2316
.
Zeng
S.
,
Guo
H.
&
Dong
X.
2019
Understanding the synergistic effect between LID facility and drainage network: with a comprehensive perspective
.
Journal of Environmental Management
246
,
849
859
.
doi:10.1016/j.jenvman.2019.06.028
.
Zhao
D.
,
Chen
J.
,
Wang
H.
&
Tong
Q.
2013
Application of a sampling based on the combined objectives of parameter identification and uncertainty analysis of an urban rainfall-runoff model
.
Journal of Irrigation and Drainage Engineering
139
(
1
),
66
74
.
(in Chinese)
.
Zheng
P.
,
Ke
J.
,
Pan
W.
,
Zhan
X.
&
Cai
Y.
2019
Effects of low-impact development on urban rainfall runoff under different rainfall characteristics
.
Polish Journal of Environmental Studies
28
(
2
),
771
783
.
doi:10.15244/pjoes/85348
.
Zhu
Z.
&
Chen
X.
2017
Evaluating the effects of low impact development practices on urban flooding under different rainfall intensities
.
Water
9
(
5487
).
doi:10.3390/w9070548
.
Zhu
Z.
,
Chen
Z.
,
Chen
X.
&
He
P.
2016
Approach for evaluating inundation risks in urban drainage systems
.
Science of the Total Environment
553
,
1
12
.
doi:10.1016/j.scitotenv.2016.02.025
.
Zhu
Z.
,
Chen
Z.
,
Chen
X.
&
Yu
G.
2019
An assessment of the hydrologic effectiveness of low impact development (LID) practices for managing runoff with different objectives
.
Journal of Environmental Management
231
,
504
514
.
doi:10.1016/j.jenvman.2018.10.046
.