Design and Evaluation of Green Space In Situ Rainwater Regulation and Storage Systems for Combating Extreme Rainfall Events: Design of Shanghai Gongkang Green Space to Adapt to Climate Change

Abstract

Global climate change has led to more extreme rainfall events. Exploring the different design schemes of rainwater in situ regulation and storage systems in green spaces to cope with extreme rainfall events is critical to cities for combating flood disasters. Using the Gongkang green space as the research object and the XPDrainage software program as the simulation tool, this study explored and evaluated different design schemes of rainwater in situ regulation and storage systems in green spaces and their responses to extreme rainfall events in Shanghai. Based on the simulated results of the runoff curves, paths and ponding area of the Gongkang green space, the ideal number and position of rainwater regulation and storage facilities were determined. Four different schemes were examined: Scheme A (diversion-oriented), Scheme B (infiltration- and detention-oriented) and schemes C and D (comprehensive rainwater regulation and storage systems). From the simulation evaluation, the total runoff volume capture rate of Scheme C reached 100%, 99.8% and 98.2% under one-, three- and five-year return period rainfall events, respectively. For the 210 mm extraordinary rainstorm event, Scheme C’s and Scheme D’s total runoff volume capture rates reached 81.9% and 94.7%, respectively. Therefore, the comprehensive rainwater regulation and storage schemes (schemes C and D) met the runoff control requirements under extreme rainfall events in the Gongkang green space. This study provides a technical reference for the optimal design of a rainwater in situ regulation and storage system in a green space and promotes the construction of resilient cities.

1. Introduction

With the aggravation of climate change, urban flooding caused by extreme rainfall has become one of the most common meteorological disasters. Further, the control of urban surface runoff has attracted significant attention [1]. As the main natural space of a city, urban green space is an important place for the in situ control of rainwater [2]. However, during the design of urban green spaces, designers usually focus on the rainwater discharge in the green space and ignore its function of regulating and storing runoff. Because of the entry of and trampling by tourists, soil infiltration is weakened, and the functions of the green space for rainwater retention, infiltration and storage are reduced. Most green spaces have limited abilities to regulate and store rainwater at the source [3]. Therefore, exploring and optimising the design and evaluation method of rainwater source regulation and the storage function of green spaces is important for rainwater and flood control across an entire city.

Quantitative simulation and runoff analysis are critical in designing and evaluating green space rainwater storage systems. Many hydrological and hydraulic simulation models and software programs apply to different scales, including basin, urban and local scales. Representative software/models relevant to the basin scale include BASINS, Watershed Model System (WMS), TOPMODEL and Catchment Hydrological Modeling, which focus on watershed division, regional surface runoff, flood prediction, and diffusion of non-point source pollution [4,5,6]. The hydrological and hydraulic models suitable for the urban scale include SWMM, InfoWorks ICM and MUSIC. They are mainly used in hydrological and hydraulic planning and design for urban areas or communities [7,8,9]. Although urban scale models such as SWMM and MUSIC can estimate the demand for low impact development facilities in a given region, less consideration has been given to the location and layout of facilities. The relevant results are difficult for designers to use directly [10].

The local scale hydrological model is most closely integrated with design and practice and can better support the design of small-scale rainwater and flood regulation and storage facilities. The local scale hydrological models include RECARGA, SUSTAIN and XPDrainage, and are used for in situ rainwater storage and regulation [11,12,13]. RECARGA can estimate the effects of low impact development facilities on runoff reduction, groundwater infiltration recharge, ponding time and total treated water volume [14]. SUSTAIN can be used for hydrological efficiency analysis of low impact development facilities and planning the spatial location of low impact facilities [15]. XPDrainage software can create a more detailed simulation and design of the effect of rainwater and flood regulation in low impact development facilities in local areas. It (1) supports the rapid establishment of a surface 2D model with a digital elevation model, (2) determines the rainfall-runoff path, ponding area and the best location of drainage facilities, and (3) estimates the runoff volume and water quality and their change processes with time [16]. XPDrainage software is simpler and more effective than other models for designing rainwater in situ regulation and storage systems in local areas [17].

Shanghai is a global super metropolis in the Yangtze River Delta, a plain with well-developed river networks. Due to the large proportion of impervious surfaces, flat terrain and high groundwater levels in Shanghai, there are problems associated with rapid runoff generation, difficult infiltration, and discharge. Facing the serious threat of rain and flood disasters, Shanghai urgently needs to improve its in situ treatment capacity for rainwater. Regarding the in situ regulation of rainwater, urban green space is the most important potential space for improvement and optimisation. The present study utilised the Shanghai Gongkang green space as the research object and XPDrainage as the simulation tool to explore the quantitative design and evaluation methods of the Shanghai green space’s rainwater in situ regulation and storage system. This study provides technical support for improving the rainwater in situ regulation and storage of green space in Shanghai and similar areas.

2. Study Area and Simulation Software

2.1. Study Area

The Gongkang green space is located in Gongkang village, Zhabei District, Shanghai (Figure 1). It was built in 1990 and is an important leisure and recreation area for residents. The total area is 9366.1 m²: 6850.5 m² (73.1%) of green space, 1133.1 m² (12.1%) of paved area, 1317.3 m² (14.1%) is a pond and 65.3 m² (0.7%) of buildings.

Cement walls bound the west and south sides of the Gongkang green space, the east side is directly connected to houses, and the north side is an ecological detention tank. The Gongkang green space is an independent catchment area. There is a drainage pipe in the northern pond, the only rainwater drainage channel from the green space. There are two garden roads in the green space, connecting the north-south and east-west directions. The garden road in the north-south direction is slightly lower and has become an important channel for the pooling of rainwater in the green space to the pond. The Gongkang green space is generally flat, and the surface runoff flows slowly to the pond; therefore, the natural drainage conditions in the green space are poor. The surface soil that tourists often trample has low rainwater retention and infiltration capacity. Therefore, the Gongkang green space is vulnerable to ponding, which affects community entertainment and leisure activities, and redesigning of the in situ rainwater regulation and storage system is urgently required to solve this ponding problem.

2.2. XPDrainage Software and Parameter Selection

XPDrainage is a sustainable drainage design software program developed by XP solutions. This program allows for simulation of the current rainwater and flooding situation at a site, on-site design of the LID rainwater regulation and storage system, water quality estimation, rainwater regulation and storage process simulation, and comprehensive evaluation of a design scheme’s effect on rainwater regulation and storage. XPDrainage integrates CAD to shorten the design time. Compared with other hydrological simulation software programs, XPDrainage provides a fast method to identify the current situation of runoff and the layout of the regulation and storage facilities in order to evaluate the effect of the design scheme and make the design process more reasonable and effective.

The routine workflow of rainwater regulation design using XPDrainage software includes target setting, runoff and ponding simulation in the existing site, design scheme formulation (determination of the number and location of facilities), simulation and comparison of different schemes, final scheme determination and evaluation.

The parameters required for the XPDrainage simulation include extreme rainfall curves, the site situation and parameters for regulation and storage facilities. The site situation mainly includes information regarding the surface elevation and soil permeability. Regulation and storage facility parameters include the facility structure, soil conditions and evaporation. The specific parameters and conditions for the present study were set as follows:

(1)

Extreme rainfall events in Shanghai

According to the Standard for the Design of Residential Green Space in China [18], residential green space should be able to handle rainfall with a return period of three to five years. In addition, the frequent rainfall events (rainfall with a one-year return period) and the strongest rainfall event in the study area deserve more attention. Therefore, the extreme rainfall events of one-, three- and five-year return periods and the strongest rainstorm in Shanghai were selected as the experimental extreme rainfall events. The rainfall curves (Figure 2) of different return periods were formulated by referring to a previous study on the rainfall data of Shanghai in the last 67 years [19]. The rainfall depths of the rainfall events of one-, three- and five-year return periods and the strongest rainstorm were 104 mm, 146 mm, 166 mm and 210 mm, respectively.

(2)

Site parameters

The site parameters were the surface elevation, base infiltration rate, hydraulic conductivity and evapotranspiration, and their values were obtained from a previous study (

Table 1. Site parameters in the Gongkang green space.

Parameter	Value
Surface elevation	DEM
Base infiltration rate	5.37 m/d
Hydraulic conductivity	0.80 m/d
Evapotranspiration	2.95 mm/d

) [20].

(3)

Regulation and storage facility parameters

The regulation and storage facilities included in the present study were a rainwater garden, grassed swale, green roof and permeable pavement. The structural and functional parameters of the rainwater garden, grassed swale and green roof were obtained from the relevant research results of the ecological planning and design team of Shanghai Jiaotong University [21,22,23]. In contrast, the porous pavement parameters were the permeable pavement brick and pavement slab standards [24]. See Appendix A for specific parameters.

3. Design Schemes and Evaluation Method of Rainwater Regulation and Storage in the Gongkang Green Space

3.1. Design Goals

Low impact development facilities with rainwater in situ regulation and storage functions should be designed for this space to alleviate the ponding problem associated with extreme rainfall events in the Gongkang green space. The total runoff volume capture rate most used during rainwater management was utilised to guide the design and evaluation. The surface runoff overflow in natural ecosystems is usually 10–15%, and the total runoff volume capture rate of natural ecosystems is 85–90% [25]. When considering the total runoff volume capture rate of natural ecosystems, the design goal was set as that of the total runoff volume capture rate of the Gongkang green space during extreme rainfall events exceeding 85%.

3.2. Simulation of Runoff and Ponding Area during the Early Design Stage

The XPDrainage software program was used to simulate the runoff and ponding area distribution of the Gongkang green space under different extreme rainfall events in order to estimate the number and spatial distribution of low impact development facilities. According to the XPDrainage software simulation results (Figure 3), runoff values of 246.8 m³, 347.6 m³, 394.4 m³ and 571.5 m³ would be generated under the extreme rainfall events of one-, three- and five-year return periods and the strongest rainstorm event, respectively. According to the simulated catchment paths and ponding area (Figure 4), the main catchment area was around the North-South Garden Road, which would become the location of low impact development rainwater regulation facilities.

3.3. Design Schemes

After analysing the ponding area, runoff path and site situations of the Gongkang green space, it was found that low impact development facilities of rainwater gardens, grass swales, green roofs, permeable pavements, and detention tanks can control the surface runoff sand and alleviate ponding in the Gongkang green space. To find the best scheme through XPDrainage simulation and comparison, we established four rainwater regulation and storage system schemes according to different function orientations. Scheme A was diversion-oriented, Scheme B was infiltration- and detention-oriented, and schemes C and D were comprehensive rainwater regulation and storage systems. Regarding the surface runoff paths and distribution of the ponding area in the Gongkang green space, the number and positions of the low impact development facilities for each scheme were determined. The details are as follows:

(1)

Scheme A—diversion-oriented rainwater regulation and storage system

The diversion-oriented rainwater regulation and storage system included two grass swales and an ecological detention tank (Figure 5) and was the least expensive scheme. The runoff flowed naturally into the grass swale and then entered the ecological detention tank. The site had to have suitable spaces to arrange the diversion path of the grass swale. Meanwhile, the ponding areas had to be relatively scattered, and the rapid runoff did not exceed the volume of the grass swale so that the grass swale could gradually divert the runoff. Considering the site conditions of the Gongkang green space, the grass swales were arranged along the garden road to guide the runoff. The total length of the grass swales was 130 m, the width was 1.5–2 m, and the total area was 230 m². The plants in the grass swales included Canna indica, Iris tectorum, Lythrum salicaria, Arundo donax, Reineckia carnea, Buddleja lindleyana, etc. The area of the ecological detention tank was about 1317.3 m². Aquatic plants planted in the pond included Acorus calamus, Zizania aquatica, and the like.

(2)

Scheme B—infiltration- and detention-oriented rainwater regulation and storage system

The infiltration- and detention-oriented rainwater regulation and storage system included three green roofs, five rainwater gardens, two permeable pavements and an ecological detention tank (Figure 6). All the facilities were arranged at the main catchment points in the simulation results of the XPDrainage software program. The infiltration- and detention-oriented rainwater regulation and storage system required the facilities to be targeted in the ponding area, which effectively solved the ponding problem during the initial rainfall stage. The green roofs were located on the Northwest Buildings, and had an area of about 65 m². The total area of the five rainwater gardens was 203 m², and the plants in the rainwater garden were the same as those in the grass swales. The permeable pavement was set on two main garden roads in the horizontal and vertical directions in the centre of the green space, and had a total length of 145 m, a width of 1.5 m and an area of about 217.5 m². The material was permeable brick, and the infiltration rate was 10 m/d. However, the ability of the infiltration- and detention-oriented rainwater regulation and storage system to manage continuous rainfall for a long period of heavy rainfall in a short time may not be adequate due to the limited runoff infiltration rate and the volume of the regulation and storage facilities.

(3)

Schemes C and D—comprehensive rainwater regulation and storage system

Two schemes (schemes C and D) for comprehensive rainwater regulation and storage systems were designed (Figure 7). The two schemes had three green roofs, one grass swale, two permeable pavement roads and an ecological detention tank. In addition, Scheme C had five rainwater gardens (Raingarden Nos. 1–5), whereas Scheme D had only two rainwater gardens (Raingarden Nos. 6 and 7). In the comprehensive rainwater regulation and storage system, rainwater first infiltrated. Then, the overflow part of the runoff entered the grass swale, which led to the rainwater gardens and ecological detention tank. The comprehensive rainwater regulation and storage system had high requirements for the site conditions and a large construction workload. According to the site conditions of the Gongkang green space and the simulation results obtained from the XPDrainage software program, in the comprehensive schemes, the grass swales were arranged along the garden road to divert runoff. Further, rainwater regulation and storage facilities were arranged at the main catchment points to absorb the nearby runoff. Scheme C included 230 m² of a grass swale, 203 m² of rainwater gardens, 65 m² of green roofs, 217.5 m² of permeable pavement and 1317.3 m² of an ecological detention tank. Scheme D had two more rainwater gardens than Scheme C, which had an area of 57 m².

3.4. Evaluation Method

Two methods were used to evaluate the rainwater regulation and storage design schemes in the Gongkang green space. First, each scheme’s total runoff volume capture rate under different extreme rainfall events was simulated and compared using the XPDrainage software program. Then, the total runoff volume capture rate and overflow occurrence time of the reconstructed green space were detected to compare and analyse the reliability of the XPDrainage software.

(1)

Evaluation using the XPDrainage software simulation

The four design schemes of the rainwater regulation and storage systems were introduced into the XPDrainage software program to simulate the total runoff volume capture rate of each scheme under rainfall standards of 104 mm (one year), 146 mm (three years), 166 mm (five years) and 210 mm (extraordinary rainstorm).

(2)

Field test evaluation after green space reconstruction

Based on the simulation evaluation results obtained from the XPDrainage software program, Scheme D was selected to reconstruct the Gongkang green space (Figure 8). Nine months after completing the Gongkang green space reconstruction, the total runoff volume capture rate and overflow occurrence time of the rainwater regulation and storage system in the green space were tested and evaluated.

Nine months after the reconstruction of the green space and based on the weather forecast, it was found that 15 and 16 September 2016 were when the extraordinary rainfall would occur. Therefore, a field evaluation experiment was undertaken over these two days. Hourly rainfall values were recorded using a rain sensor (Davis rain sensor S-RGC-M002, Onset HOBO, Bourne, MA, USA). The sensor was installed in the Gongkang green space the day before the rainfall based on the weather forecast in order to record the continuous rainfall for 48–72 h.

The times when the runoff overflowed the facilities were manually recorded. The runoff accumulation at each facility was observed every 5 min after rainfall, and the overflow start time of each facility was recorded.

The recorded rainfall curves were imported into the XPDrainage software program. The overflow time of each facility was simulated and compared with the field test results in order to evaluate the reliability of the XPDrainage software program.

4. Evaluation Results

4.1. Evaluation of the Diversion-Oriented Rainwater Regulation and Storage System (Scheme A) Using XPDrainage Simulation

Based on the simulation results (Figure 9), the rainwater control rates of Scheme A for one-, three-, and five-year return periods and the 210 mm rainfall event were 99.2%, 81.9%, 72.0% and 48.0%, respectively. Under the rainfall events of the 1- and 3-year return periods, Scheme A better controlled the runoff generation. Under the 5-year return period and 210 mm rainfall event, a large amount of runoff was generated in the green space, and the conduction capacity of the grass swale to runoff was insufficient. To improve the regulation and storage capacity of the green space, it was necessary to increase the number and depth of the grass swale. However, due to the limitation of the terrain, there was no other path for the green space to guide the rainwater to the ecological detention tank, and the depth of the grass swale should not be increased; therefore, the regulation and storage capacity of the system could not be effectively improved. Thus, Scheme A could not meet the requirements of the Gongkang green space to control the runoff of extreme rainfall events.

4.2. Evaluation of Infiltration- and Detention-Oriented Rainwater Regulation and Storage System (Scheme B) Using XPDrainage Simulation

Based on the simulation results (Figure 10), the rainwater control rates of Scheme B for one-, three-, and five-year return periods and the 210 mm rainfall event were 84.8%, 80.3%, 77.0% and 65.6%, respectively. Scheme B had a limited ability to control runoff under extreme rainfall events and failed to achieve the design goal. When the volume of rainwater exceeded the storage capacity of the facilities, the facilities could only rely on the infiltration to slowly treat the rainwater, which had a poor ability to respond to continuous rainfall. However, due to the different amounts of rainwater obtained by the facilities and the lower connection between the facilities, overflow occurred in some areas first; thus, it was difficult to reach the entire system’s maximum regulation and storage capacity. The terrain of the Gongkang green space is flat, and the utilisation efficiency of the ecological detention tank in the green space is low in the absence of diversion facilities. Therefore, the suitability of this independent infiltration- and detention-oriented rainwater regulation and storage system (Scheme B) in Gongkang green space was poor.

4.3. Evaluation of Comprehensive Rainwater Regulation and Storage Systems (Schemes C and D) Using XPDrainage Simulation

Based on the simulation results of Scheme C (Figure 11), this system had good control of runoff, reaching a 100% total runoff volume capture rate under the one-year return period rainfall event, and 99.8% and 98.2% total runoff volume capture rates under the three- and five-year return period rainfall events, respectively. However, under the extreme 210 mm rainstorm event, the total runoff volume capture rate of the green space was 81.9%, which did not meet the design requirements.

Compared with Scheme C, Scheme D had two rainwater gardens (No. 6 and No. 7) with a stronger treatment capacity for rainwater runoff. Based on the simulation results of Scheme D (Figure 12), the total runoff volume capture rate in the green space during the 210 mm rainstorm event increased from 81.9% (in Scheme C) to 94.7%, which met the design requirements of the Gongkang green space.

4.4. Field Evaluation of Scheme D and Reliability Analysis of the XPDrainage Software Program

Scheme D was used to reconstruct the Gongkang green space, and the field test and evaluation were undertaken on 15 and 16 September 2016, nine months after the reconstruction. The experiment recorded the rainfall every 5 min from 18:00 on 15 September 2016 to 18:00 on 16 September 2016, and the daily rainfall was 196.9 mm. The facilities controlled the runoff of the Gongkang green space at 662.3 m³ and 642.5 m³, which met the 87.1% total runoff volume capture rate and achieved the design goal.

The daily rainfall pattern was introduced into the XPDrainage software program. The difference between the simulated overflow time and the measured time of each rainwater garden was compared to evaluate its reliability (

Table 2. Overflow occurrence time of rainwater gardens in the Gongkang green space.

Rain Garden No	Simulated Overflow Time (Min)	Measured Overflow Occurrence Time (Min)
No. 1	-	-
No. 2	-	-
No. 3	940	895
No. 4	-	-
No. 5	1280	1020
No. 6	-	-
No. 7	-	1365

).

As shown in Table 2, the accuracy of the XPDrainage software program for predicting the occurrence time of the facility overflow was generally good. In Nos. 3 and 5 rainwater gardens, the field observed overflow occurrence times were 45 min and 160 min earlier than the simulated occurrence time, with 5% and 15% errors, respectively. The possible reason for this was that the rainwater garden was in the state of natural water content daily, larger than that set by the XPDrainage software program, resulting in the actual runoff generation time being earlier than that simulated. Only the simulated overflow occurrence time of the No. 7 rainwater garden was quite different from the field observed one. The simulation results showed that the No. 7 rainwater garden did not overflow; however, the measured overflow occurred in 1365 min. This discrepancy might be because the terrain of the No. 7 rainwater garden was lower than that of the other rainwater gardens, and the overflow runoff from the upstream rainwater garden infiltrated into the No. 7 rainwater garden after collection, resulting in an overflow.

Overall, four of the seven rainwater gardens did not overflow, which was consistent with the simulation, and two of them overflowed earlier than in the simulation. Only one of the rain gardens that was thought not to overflow in the simulation did overflow. The XPDrainage software program performed well in simulated runoff generation.

5. Discussion and Conclusions

The runoff curves, paths and ponding areas of the Shanghai Gongkang green space in the extreme rainfall events of one-, three-, and five-year return periods and an extraordinary 210 mm rainfall event were simulated using the XPDrainage software program. Four different schemes (diversion-oriented, infiltration- and detention-oriented and comprehensive rainwater regulation and storage systems) for combating extreme rainfall events in the Gongkang green space were designed based on the simulation results. From simulation evaluation, it was found that the total runoff volume capture rate of Scheme C reached 100%, 99.8% and 98.2% under 1-year, 3-year and 5-year return period rainfall conditions. For the 210 mm rainstorm event, Scheme C’s and Scheme D’s total runoff volume capture rates reached 81.9% and 94.7%, respectively. Therefore, the comprehensive rainwater regulation and storage scheme met the runoff control requirements under extreme rainfall events in the Gongkang green space. In general, this study used XPDrainage simulation to find a nature-based solution design scheme for the Gongkang green space to deal with extreme climate.

In rainwater and flood regulation and storage design, basin scale and urban scale hydrological simulation software, such as BASINS, WMS, SWMM and MUSIC, are used more frequently. Meanwhile, the local scale software XPDrainage is used less frequently [6,7,8]. However, some studies have proved that XPDrainage software has more advantages than basin and urban scale software in accurately determining the runoff path, ponding area and the best location of regulation and storage facilities [16,17]. This study also proves that XPDrainage software can accurately identify the runoff path and ponding area in the Gongkang green space and has high accuracy in simulating the location and time of overflow. This is of great help to sponge city design with the goal of in situ rainwater regulation and storage.

This study found that the runoff generation simulated by XPDrainage software may occur later than the actual runoff generation. The reason for this may be that Shanghai is a plain water network area with high groundwater levels, high soil natural water content and short runoff generation times. Therefore, special attention should be paid to this error when using XPDrainage software for simulation and design under similar environmental conditions.

Generally, the XPDrainage software program is suitable for designing and evaluating green rainwater in situ regulation and storage systems. It made the design process more convenient and effective, and provided a reference for guiding the design and evaluation of green facilities for extreme rainfall events in response to climate change in Shanghai.