Multi-Level Control and Utilization of Stormwater Runoff

Abstract

This study proposes the technology of “runoff storage and seepage utilization” for achieving purification of road rainfall–runoff and presents a multi-level series purification system (PBT-GR) comprising porous asphalt pavement (PAP), a bioretention system (BS), a storage tank (T) and a hydroponic green roof (GR). The operation parameters of each component unit were optimized and the contribution of each unit to pollution was analyzed. The results showed that under typical simulated rainfall, the suspended solids (SS), total nitrogen (TN), total phosphorus (TP), Pb, Zn and Cu removal rates by filtration and interception of porous pavement were 62.26 ± 3.19%, 16.29 ± 1.74%, 29.27 ± 1.37%, 37.61 ± 2.58%, 35.57 ± 4.64% and 31.17 ± 3.27%, respectively. The average concentrations of SS, TN, TP, Pb, Zn and Cu in the effluent of the PBT-GR system were 14.70 ± 2.21 mg/L, 1.52 ± 0.24 mg/L, 0.14 ± 0.04 mg/L, 0.09 ± 0.04 mg/L, 0.11 ± 0.03 mg/L and 0.04 ± 0.01mg/L, respectively, which met the water quality standards recommended in the Chinese guidelines and showed a high adaptability to pollution load. The contents of pesticide residues and heavy metals in cultivated vegetables met the national standards. The period required to recoup the investment in the system was approximately 3 years, indicating its good economic feasibility. The present study can provide a valuable reference of the construction of an efficient, low consumption and sustainable urban stormwater treatment system and can contribute to the improvement in the quality of the urban water environment.

1. Introduction

The rapid acceleration of urbanization has resulted in large areas of surface water area, grassland and other natural permeable ground being gradually replaced by impervious pavement, which has consequently resulted in a series of environment problems including deterioration of the urban water environment, frequent flood disasters and serious pollution of rainfall–runoff [1,2,3]. Urban non-point source pollution is one of the most important drivers of degradation of urban water quality. Roads are an important component of the urban catchment surface, and since they are always closely connected with drainage facilities, they are a key source of urban non-point source pollution. Source-based, decentralized, small-scale and low-impact development (LID) rainwater control measures have been piloted across China to effectively address the challenges presented by urban road stormwater systems. Studies in both China and abroad [4,5,6] show that regardless of whether centralized infiltration technology, such as infiltration pool/pond, porous pavement and rainwater gardens, or horizontal transmission technology such as grass ditches or reservoirs are used, the high accumulation of nitrogen, phosphorus and heavy metals by the long-term operation of technology alone or in combination poses serious pollution risks to surface water, groundwater and soil [7,8]. Integration of urban stormwater management into agriculture and forestry could facilitate the transformation of problematic nitrogen, phosphorus and metal elements in road stormwater runoff into valuable fertilizer resources. However, the agricultural and forestry industries require large areas of land, whereas urban land resources are extremely limited. Within this context, the present study proposes the concept of combining a stormwater treatment unit comprising a porous pavement combined with bioretention ponds with a stormwater utilization unit comprising a green aquatic vegetable roof, which promotes sustainable development through stormwater recycling.

A hydroponic green roof (GR) is an ecological landscape garden constructed on a roof, terrace or balcony. GR is an effective technology for reducing rainfall–runoff from the source [9,10,11] and is usually used for rainwater collection, storage and treatment. However, traditional green roofs are associated with various disadvantages such as excessive load bearing, limited rainwater-runoff interception capacity and the need to irrigate during the dry season [12]. The improvement in current methods of stormwater storage and purification is a component of stormwater resource utilization technology. A hydroponic vegetable wetland has strong nitrogen and phosphorus removal capacity [13,14,15]. Water treated by a combined LID facility can provide a nutrient solution for a hydroponic green roof. The large root system of a hydroponic green roof can absorb nitrogen, phosphorus and some other water quality variables for plant growth [16]. Income can be gained from a green roof by harvesting vegetables, thereby realizing the utilization of nitrogen and phosphorus [17,18,19]. The use of a horizontal vegetable wetland on top of an urban building roof can not only relieve shortages of urban land, but also facilitate the purification of road rainfall–runoff with the additional economic benefits.

The present study developed a system (PBT-GR) comprised of porous asphalt pavement (PAP), a bioretention system (BS), a storage tank (T) and a hydroponic green roof (GR), and the relevant parameters of segmented units were optimized. The contributions of each unit to the removal of different pollutants were determined through changes in pollutant concentrations throughout the treatment period. The present study also discusses the benefits of the system from the perspectives of water quality safety, the economic feasibility of the system and the safety of vegetables produced through the system for human consumption. The PBT-GR system integrates a green roof with a stormwater ecological treatment unit. The addition of a small amount of light filler into a green roof can not only effectively reduce the bulk density of the roof, but also enable additional water storage by the system. The PBT-GR system can realize the collection, treatment and storage of natural rainwater, as well as the treatment of road rainfall–runoff. In addition, the PBT-GR system can provide economic benefits and can help improve the urban water ecological environment. Therefore, the results of the present study can provide a valuable reference for the sustainable utilization of water resources and the improvement in the urban living environment.

2. Materials and Methods

2.1. Experimental Apparatus

The PBT-GR system mainly comprises a porous asphalt pavement, a bioretention system, a storage tank and a green roof for the hydroponic growth of vegetables. The stormwater produced by a simulated rainfall system firstly enters the porous asphalt pavement, following which the water enters the bioretention system through gravity, then flows into the storage tank and is finally transmitted to the hydroponic green roof through a peristaltic pump. The porous asphalt pavement used was an I-type semi-permeable pavement with a size of 300 mm (length) × 300 mm (width) × 200 mm (height) and a void ratio of 24.3%. All structural layers were designed in accordance with relevant domestic standards, such as the CJJ/T 190-2012 standard of porous asphalt pavement [20]. The bioretention system comprised a cylinder composed of organic glass with a catchment area of 1:10, a diameter of 60 mm and a height of 600 mm. The bioretention system incorporated a submerged layer (100 mm), filter medium layer (500 mm) and gravel drainage layer (100 mm). The filter medium layer was composed of a uniform mixture of 80% natural sand and 20% soil mass, whereas the drainage layer was composed of gravel with particle size of 5 mm–10 mm. The dimensions of the hydroponic green roof were 1200 mm (length) × 300 mm (width) × 400 mm (height) and it was composed of a light substrate layer of 200 mm thickness containing water spinach. Figure 1 is a photograph showing the test device.

2.2. Experimental Design

Several simulated rainfall tests, including rainfall intensity and infiltration concentration, were conducted to determine the efficiency of the PBT-GR system for reducing and utilizing urban rainwater under complex conditions. Three different rainfall recurrence periods of one year (1a), three year (3a) and five year (5a) were designed to investigate variation in rainfall intensity. The pollutant concentrations of the artificial runoff were set according to measurements of the same pollutants in road surface rainfall–runoff in Jiangsu province in China, and two different levels of pollutants in artificial runoff were set in the experiment (

Table 1. Levels of pollutant concentrations used in artificial runoff in the study (mean ± standard deviation, n = 3).

Pollutant	SS (mg·L−1)	NH4+-N (mg·L−1)	TN (mg·L−1)	TP (mg·L−1)	Pb (mg·L−1)	Zn (mg·L−1)
Typical concentration	200 ± 10	3 ± 0.3	6 ± 1	0.8 ± 0.2	0.5 ± 0.1	1 ± 0.2
High concentration	400 ± 20	6 ± 0.6	12 ± 2	1.6 ± 0.4	1 ± 0.2	2 ± 0.4
Reagent	Kaolin	NH4Cl	NH4Cl/KNO3	KH2PO4	PbCl2	ZnCl2

Abbreviations: SS, suspended solids; TN, total nitrogen; TP, total phosphorus.

). The current study designed a rainfall process with a duration time of 120 min based on an empirical model of rainfall intensity in Nanjing and the Chicago rainfall model [21].

Polyethylene bottles were used to collect rainfall–runoff samples from the porous asphalt pavement (sampling point A), the bioretention outlet (sampling point B), the storage tank (sampling point C) and the hydroponics green roof (sampling points D and E). Samples were obtained as soon as rainfall–runoff was produced by the porous asphalt pavement. Samples in point A of rainfall–runoff from the pavement were collected at 10 min, 30 min, 60 min, 90 min and 120 min. Samples in point B were collected from the bioretention outlet at 30 min, 60 min, 90 min and 120 min according to the catchment volume. After sample collection, the concentrations of suspended solids (SS), NH₄⁺-N, total nitrogen (TN), total phosphorus (TP), Pb and Zn in the runoff were measured.

2.3. Data Analysis Methods

2.3.1. Efficiency Assessment

There were large variations in the concentrations of pollutant in surface runoff over a rainfall event due to the influence of rainfall characteristics and pollutant properties. This present study adopted the event mean concentration (EMC) based on the experience of the United States Environmental Protection Agency (USEPA) [22] as a measure of pollutant concentration. The calculation of the EMC was a measure of pollutant concentration. The calculation of the EMC was as follows:

$$EMC=\frac{M}{V}=\frac{{\int }_0^T{c}_t{Q}_{t}dt}{{\int }_0^T{Q}_{t}dt}\cong \frac{\sum c_t{Q}_t\Delta t}{\sum Q_t\Delta t}$$

(1)

In Equation (1), EMC is the event mean concentration of runoff pollutant (mg/L), M is the load of pollutant entering the system over the entire runoff process (mg), V is the total runoff volume (L), c_t is the concentration of a pollutant at time T (mg/L), Qt is runoff flow at time t (L/ min) and Δt is the sampling interval (min).

Storm runoff pollution load refers to the total removal load discharged by runoff due to one or more rainfall events. The runoff pollution load per unit area of a pollutant can be expressed as follows:

$$L=0.001\sum _{i=1}^{m}EM{C}_i\times R_i\times P_i$$

(2)

$$R_c=Q\times \left[\sum _{i=1}^n(C_{oi}-C_{ei})\times t_i\right]$$

(3)

In Equations (2) and (3), L is the runoff pollution load per unit area of a pollutant (mg·m⁻²), 0.001 is the unit conversion factor, m is the number of samples, R_i is the runoff coefficient of the ith rainfall event, R_i of the porous pavement was assigned a value of 0.8 based on experience [23], P_i is the rainfall capacity of the ith rainfall event (mm), R_c is the load of pollutant removed [g·(m²·d)], C_oi and C_ei are the average inlet and outlet concentrations (mg·L⁻¹), respectively and t is the time interval between the sampling time (d).

Removal rates (RR) of the major parameters of SS, TN, NH₄⁺-N, TP, Pb and Zn were computed by the formulae given below:

$$RR=\frac{C_i-C_e}{C_i}$$

(4)

In Equation (4), C_i and C_e are the inflow and outflow concentrations in mg·L⁻¹.

Hydraulic Loading Rates HLR (L∙m⁻²∙d⁻¹) applied to the hydroponic green roof were determined using the equation:

$$HLR=\frac{Q}{A_w}$$

(5)

In Equation (5), Q is the flow (L∙d⁻¹) of stormwater runoff through the hydroponic green roof, A_w is the area (m²) of the given experimental pond of the constructed wetland.

2.3.2. Statistical Analysis

Statistical analysis was performed using IBM SPSS version 20.0 to analyze the performance of PBT-GR under different operating conditions for significance with p < 0.05 and the Pearson coefficient for correlation analysis.

3. Results and Discussion

3.1. The Effect of Hydraulic Loading Rate (HLR) on Pollutant Removal by the Hydroponic Green Roof

As shown in Figure 2a, the TN removal rate generally decreased with increasing HLR. The TN removal rate was highest at an HLR of 0.1 m³·(m²·d)⁻¹, reaching 63.2%, and was lowest (only 31.7%) when HLR increased to 0.4 m³·(m²·d)⁻¹. The reason for this observation was that the increase in HLR resulted in a decrease in the hydraulic retention time, leading to a short contact time between the substrate hosting microorganisms and pollutants and consequently reducing the treatment effectiveness. As HLR increased, the removal load first increased and then decreased. The removal load was lowest at an HLR of 0.1 m³·(m²·d)⁻¹ at 0.054 g·(m²·d)⁻¹. The removal load reached a maximum when HLR was 0.3 m³·(m²·d)⁻¹ at 0.135 g·(m²·d)⁻¹. The removal load reduced to 0.011 g·(m²·d)⁻¹ when HLR increased to 0.4 m³·(m²·d)⁻¹. These results indicate that increasing HLR can increase the removal load within a certain range. However, when the HLR exceeds a certain value, dissolved pollutants in rainwater cannot be effectively degraded due to excessive water flow velocity. The changes observed in the TP removal rate were similar to that of the TN removal rate (Figure 2b). The TP removal rate first increased and then decreased with an increase in HLR (p < 0.1). The removal load reached a maximum of 0.024 g·(m²·d)⁻¹ when HLR was 0.3 m³·(m²·d)⁻¹.

The metal removal rate reached a maximum above 65% when HLR was 0.1 m³·(m²·d)⁻¹, following which the metal removal rate gradually decreased. The metal removal rate was lowest below 50% when HLR was 0.4 m³·(m²·d)⁻¹. These results indicate that a low HLR was conducive to the adsorption of heavy metal ions on the substrate surface. Pb²⁺ and Zn²⁺ in runoff were precipitated as Pb(OH)₂ and Zn(OH)₂ and adsorbed onto the substrate surface. They may also have reacted with CO₃²⁻ to form heavy metal carbonate precipitation complexes [24].

$${\mathrm{Pb}}^{2+}+{\mathrm{CO}}_3^{-2}\iff P\mathrm{b}{({\mathrm{CO}}_3)}_2$$

(6)

$$\mathrm{Zn}+{\mathrm{CO}}_3^{-2}\iff \mathrm{Zn}{({\mathrm{CO}}_3)}_2$$

(7)

The Pb removal rate gradually increased with increasing HLR (p < 0.1), with the maximum load of Pb removed being 0.012 g·(m²·d)⁻¹. The load of Zn removed with the change in HLR was similar to that of TN and TP, showing first an increase followed by a decrease. The Zn removal rate reached a maximum when the HLR was 0.3 m³·(m²·d)⁻¹ at 0.028 g·(m²·d)⁻¹. As such, nutrients should be removed in as short a time as possible to achieve full utilization of nitrogen and phosphorus nutrients. The test results showed that the maximum loads of TN, TP and Zn were reached when HLR was 0.3 m³·(m²·d)⁻¹. Heavy metals were effectively treated in the initial runoff. These results indicated that the HLR of a hydroponics green roof should be 0.3 m³·(m²·d)⁻¹ to ensure the maximum utilization of nitrogen and phosphorus resources.

3.2. Effect of Rainfall Intensity on Pollutant Removal by the Porous Pavement Bioretention (PB) System

Figure 3 shows the effectiveness of the PB system in removing pollutants from stormwater runoff under three rainfall intensities. There was a gradual increase in the concentrations of TN, TP, Pb and Zn with an increasing duration of rainfall, but not for SS. The results showed that the SS removal rate within the PB system decreased with increasing rainfall intensity. The greatest SS removal rate was achieved at a rainfall intensity of 1a. The concentrations of SS in the effluents of the porous asphalt pavement and bioretention systems were 35.8 mg/L and 21.1 mg/L, respectively. However, an increase in rainfall intensity to 5a resulted in the SS removal rate from the effluent of porous asphalt pavement and bioretention tank systems decreasing by 22.5% and 16.1%, respectively. This result was consistent with research by Li Haiyan [25] and Wang Dongqi [26]. The removal of SS by the porous asphalt pavement and bioretention pond mainly relies on interspace interception and surface adsorption. Due to the high rainfall intensity and obvious scour effect under heavy rainfall, rainwater runoff will rapidly flush SS through the porous asphalt pavement and bioretention ponds. The concentration of NH₄⁺-N increased with increasing rainfall duration, thereby showing an opposite trend to that of SS. The NH₄⁺-N removal rate by the PB system decreased by 10–13% during the later period compared with that in the earlier stage of rainfall. The NH₄⁺-N removal rate gradually decreased with increasing rainfall intensity. The NH₄⁺-N concentration of the effluent of the PB system was 1.18 mg/L when the rainfall intensity was 1a, falling within the class IV standard for surface water in China. An increase in the rainfall intensity to 5a resulted in the NH₄⁺-N removal rate of the PB system decreasing by 15.8%, with the effluent NH₄⁺-N concentration reaching 1.68 mg/L. The change in TN was similar to that of NH₄⁺-N, although the NH₄⁺-N removal rate in the PB system reached >30% under a rainfall intensity of 1a, the TN removal rate reached a maximum of only 15.3%. This result showed that there was a change in nitrogen forms in the PB system without overall removal, while the bioretention unit greatly improved the TN removal rate. The inclusion of a bioretention unit can increase the TN removal rate by 27.1–32.9% compared to use of a single porous asphalt pavement unit, mainly due to its strong water retention and adsorption capacity [27]. In addition, the presence of an anoxic region in the lower part of the bioretention unit may facilitate denitrification to remove nitrate nitrogen, thereby improving the TN removal rate. These results are consistent with those of Gong et al. [9]. Some differences were evident compared to those by Jiang Wei et al. [28] mainly due to the high voidage of the porous asphalt pavement in the present study and the low pollutant concentrations of artificial rainwater [27]. The TN concentrations of the effluent in the PB system under a rainfall intensity of 1a–5a were 3.2 mg/L–3.8 mg/L, thereby exceeding the class V surface water limit level in China. The direct discharge of effluent with this TN concentration range would have serious effects on the ecological environment. Removal of nitrogen from the effluent could provide fertilizer resources needed for agriculture.

The TP removal efficiencies under different rainfall intensities remained low, ranging between 65.3–74.2%, with the treated water TP concentration ranging between 0.2 mg/L–0.28 mg/L, falling within the class IV surface water standard. Adsorption and chemical precipitation may occur on the surfaces of different media [29]. There was an increase in the concentration of phosphorus within initial rainwater with increasing rainfall duration (p < 0.1), which may be due to the gradual occupation of fast reversible adsorption points on the surfaces of media with different properties. The changes in Pb and Zn removal rates under different rainfall intensities were consistent with those of N and P, ranging between 68.1–82.6% and 50.8–68.7%, respectively under different rainfall intensities. The Pb concentration of outflow was 0.08 mg/L–0.16 mg/L, slightly exceeding the class V surface water standard (p < 0.1), thereby indicating that Pb in outflow of the PB system will exceed the standard under strong rainfall. The concentration of Zn of the PB system outflow was 0.28 mg/L–0.44 mg/L, falling well within the class II surface water standard. The initial rainwater treated by the PB system contained relatively higher nitrogen elements such as NH₄⁺-N and TN compared to heavy metals, which provides theoretical support for the design of subsequent ecological utilization units.

3.3. The Underlying Processes Driving Runoff Pollution Load Reduction by the PBT-GR System

Figure 4 shows the pollutant concentration of each unit of the combined PBT system. The removal of pollutants by the PBT-GR system mainly relied on the combined action of the porous pavement, bioretention system and the hydroponics green roof. The removal rates of different pollutants increased as water moved through the different units. Water quality achieved through secondary purification by the bioretention system improved significantly compared to that achieved by primary purification through the porous pavement. Compared to that achieved by the porous pavement, the bioretention system increased SS, N, P, Pb and Zn removal rates from 63.9 ± 2.1%, 18.8 ± 0.5%, 27.8 ± 1.2%, 36.8 ± 3.4%, 30.4 ± 2.7% to 90.5 ± 6.7%, 53.3 ± 4.3%, 73.8 ± 4.7%, 80.4 ± 7.1% and 72.1 ± 6.9%, respectively. Figure 5 shows the contributions of different components of the PBT-GR system to the overall removal of different pollutants. The removal rates of N, P and heavy metals achieved by the bioretention unit (>35%) were higher than those achieved by the porous pavement (<36.1%) and green roof (<27.1%). The green roof achieved lower removal rates compared to the bioretention unit since rainwater entering the green roof had relatively stable pollution loads. The aerobic and anaerobic environment of the green roof wetland provided an environment suitable for sustaining the growth of plants and microorganisms [30], and therefore the green roof was effective at removal of nitrogen pollutants. The removal rate of SS by the porous pavement (>63.9%) was significantly higher than those of the bioretention system (17.6%) and green roof (13.9%), respectively. This was due to the presence of unique internally connected and semi-connected cavities of the porous asphalt pavement, which formed a dense network structure. The majority of particulate pollutants were intercepted by the upper structure of the porous pavement, with only a small portion of sediment with a small sediment particle size passing through the porous asphalt pavement into the bioretention system and reservoir. This result is consistent with that obtained by Li Haiyan [25] and Cui Xinzhuang [30]. Within the entire PB system, the role of the reservoir was mainly to regulate water storage, while the removal of SS through precipitation in the reservoir was limited. This was because rainwater had a short residence time in the reservoir and because the majority of particulate matter had been intercepted by the porous asphalt pavement.

3.4. Effect of Pollutant Concentration Range on Pollutant Retention by the PBT-GR System

Figure 6 shows the effect of varying concentrations of SS, N, P and Zn on the treatment capacity of the PBT-GR system. A three-day antecedent period was used in all experiments to discount the effect in the antecedent period. Differences in the effectiveness of the combined treatment systems under normal and extreme events were analyzed for statistical significance. It is generally recognized that inflow stormwater of higher pollutant concentrations would result in higher pollutant concentrations outflow of the PBT-GR system [15,27]. The results showed that the efficiency of TP removal increased under higher concentrations (extreme event) for both the porous asphalt pavement and bioretention systems. Similar results [31] were observed by Rahman et al. (2020); a higher phosphorus load represented a larger concentration gradient, which in turn promoted adsorption of phosphorus in the solution to the substrate by enhancing available chelate sites on the adsorbent. Similarly, the PBT-GR system exhibited excellent removal efficiencies for Pb, Zn and Cu of >73.5% under a high concentration of heavy metals (p < 0.1). The PBT-GR system maintained sufficient treatment capacity to effectively treat rainwater with excessive concentration of heavy metals within a short period of time.

The PBT-GR system showed a higher TN removal efficiency under high TN concentrations (extreme event) compared to that under normal TN concentrations, with the bioretention and green roof systems performing particularly well at 59.4% and 83.1%, respectively. Lucke et al. [32] and Gong et al. [9] determined that the removal ratios of TN in a bioretention and a green roof system under high inflow concentrations were 30.1% and 72.4%, respectively, which were higher than that at low pollutant inflow concentrations. The PBT-GR system was able to simultaneously remove nutrients and heavy metals from runoff. It was inferred that the typical concentrations of SS, TN, TP, Pb and Zn in the porous asphalt pavement and bioretention system under a rainfall recurrence interval of 2a were 9.2 mg/L, 1.2 mg/L, 0.11 mg/L, 0.08 mg/L and 0.13 mg/L, respectively, whereas hydraulic loading in the green roof was 0.3 m³·(m²·d)⁻¹. Under the high pollutant concentration (extreme event) experiment, the concentrations of SS, TN, TP, Pb and Zn were found to be 13.6 mg/L, 1.94 mg/L, 0.15 mg/L, 0.14 mg/L and 0.29 mg/L, respectively. A noteworthy outcome of the present study was the observation that the PBT-GR system was enabled to simultaneously remove nutrients and heavy metals from runoff, even under extreme pollutant concentrations.

3.5. Comprehensive Analysis of the Benefits of PBT-GR

3.5.1. Water Quality Safety Assessment

The quality of the effluent of the PBT-GR system was examined to evaluate the direct impact of the effluent on the surrounding environment and explore the feasibility of recycling rainwater resources for irrigation purposes. The quality of the effluent was evaluated according to the Environmental Quality Standard for Surface Water (GB 3838-2002), [33] the Groundwater Quality Standard (GB/T 14848-2017) [33] and the Water Quality Standard for Farmland Irrigation (GB 5084-2005) [34]. Figure 7 shows the effluent quality of the PBT-GR system according to the main water quality variables assessed. The concentrations of all pollutants in the effluent of the PBT-GR system fell within the class V surface water and irrigation water standards. The effluent produced by the PBT-GR can therefore be used for landscape water, industrial and agricultural applications discounting other requirements, and the concentrations of heavy metals in the effluent did not pose a short-term toxicity risk to surrounding waters. The concentration of SS in the effluent of the PBT-GR was far below the limit of the water quality standard for farmland irrigation, whereas the TN concentration exceeded the class V surface water and groundwater limits, indicating a limited capacity of the PBT-GR system to remove N pollutants >2.8 mg/L. Therefore, the effluent of the PBT-GR system was found to be unsuitable for direct release into groundwater and surface water resources, and should only be used as irrigation water. The concentrations of Pb and Zn in the PBT-GR system were 0.1 mg/L and 0.3 mg/L, respectively, which were within the irrigation water standard and the concentration of Zn and appropriate for plant growth.

3.5.2. Evaluation of the Safety of Vegetables Produced by the Green Roof for Consumption

The vegetables produced by the green roof were assessed for heavy metal content. Samples of water spinach were obtained from different positions under different operation stages of the hydroponic green roof system. Some of the pollutants were detected according to the Test Catalogue of 55 Kinds of Pollution-Free Agricultural Products Including Eggplant and Fruit Vegetables (Nongbanzhi (2013) No. 17 document) [35] and Contaminant Limits in Food (GB 2762-2017) [36].

Table 2. Results of tests for metals in water spinach produced by the GR system (mean ± standard deviation, n = 6).

No.	Test Item	Result (mg·kg−1)		Threshold (mg·kg−1)
		Green Roof-1	Green Roof-2
1	Lead	0.07 ± 0.01	0.05 ± 0.01	0.1
2	Cadmium	0.04 ± 0.01	0.02 ± 0.01	0.1
3	Chromium	0.21 ± 0.02	0.14 ± 0.02	0.5
4	Copper	0.17 ± 0.03	0.22 ± 0.04	10
5	Tamaron	Not detected	Not detected	0.05
6	Omethoate	Not detected	Not detected	0.02
7	Dursban	Not detected	Not detected	0.05

clearly shows that the contents of the heavy metals Pb, Cd and Cr in water spinach produced with water before treatment by the PBT-GR system (GR-1) were higher than that in water spinach produced using final effluent of the PBT-GR system (GR-2), although the concentrations of these metals at both positions were well within the national standard limit, and no pesticide residues and heavy metals exceeded the standard.

3.5.3. Analysis of the Economic Benefits of the Combined System

An economic analysis of the rainwater collection system was conducted in an urban road with a catchment water of 200 m², incorporating construction cost, maintenance cost, investment payback period and revenue cost, with the results shown in

Table 3. Estimation of investment and economic benefit of the PBT-GR system.

Stage	Object	Price (Yuan)	Note	Reference
Investment cost	Underground concrete reservoir (10 m3)	2642	/	[37]
	Green roof of hydroponic wetland	3142	Wetland pool, substrate and vegetable seedlings	[9]
	Pipes, fittings, sinks	504	/	[9]
Maintenance cost (per year)	Reservoir cleaning	113	/	[38]
	Pipeline and wetland maintenance	64	/	[39]
Annual economic income	Saved irrigation water resources	27	According to the electricity consumption in the process of producing reclaimed water	[40]
	Saved fertilizer	42	Refer to nitrogen fertilizer and phosphate fertilizer	[40]
	Income from planting vegetables	2582	Based on production and local unit price	[9]

. The results showed that the system is affordable for middle and low-income residents, with low estimated maintenance costs as the system is robust and resilient to damage over a long period. In addition, the benefit estimation indicated the investment payback period to be approximately 3 years, in other words, the investment cost can be recovered within 3 years. Given the fact that water resources are becoming increasingly limited and that water may become more expensive in the near future, rainwater collection and utilization shows a great potential for achieving safe water and crops, and the approach has been shown to be economically feasibility.

3.6. Viability of the PBT-GR System

Hydroponic wetland green roofs require less planting substrate and provide a greater storage capacity compared with traditional rainwater treatment facilities. These benefits not only increase the rainwater storage capacity of the system, but also reduce the load on a roof. Because the bulk density of soil (1.5–2.5 g·cm⁻³) is greater than that of water (1 g·cm⁻³), the bulk density of the system is reduced by the use of the hydroponic method instead of the traditional soil cultivation method. At a storage height of 200 mm, the total weight of water, light substrate, facilities and plants is about 84 kg·m⁻², far lower than the load standard for green roofs of most urban buildings of 200 kg·m⁻² [41]. In addition, the water level of the hydroponic green roof can be reduced to 20 mm from the bottom, and the stormwater runoff recovery rate can reach 90% under a rainfall intensity < 100 mm·d⁻¹. Therefore, this system shows a greater potential in reducing urban floods compared with traditional green roof systems.

The PBT-GR system was shown to effectively treat road rainwater runoff, with the quality of the effluent to be similar or improved compared to that of membrane bioreactors, sequencing batch reactors and LID facilities [42,43,44]. After disinfection, the effluent would be of sufficient quality for application in the flushing of toilets, irrigating landscapes and flushing roads. In addition, the installation of ecological treatment facilities such as the PBT-GR system on roofs can help alleviate problems associated with excessive land use. However, significant differences in rainfall and water quality exist for different regions and seasons. The rainfall interval periods over different regions vary greatly, particularly for some arid areas, which can easily result in insufficient collection of rainwater or excessive supply of rainwater to a hydroponic green roof, and therefore rainwater runoff storage and utilization efficiency is dependent on regional climate. The present study attempted to explore the integration of urban domestic grey water treatment technology with a green roof hydroponic wetland based on the published domestic and international engineering application studies to meet irrigation water demand during a drought period. The feasibility of this approach was preliminarily studied with the assumptions of the roof area being 100 m² and the per-capita daily water usage being 50 L and by reference to the configuration concentration by Song et al. [15]. Figure 8 shows the results of this analysis. Under a hydraulic retention time (HRT) of 3 days, the effluent concentrations of COD (Chemical Oxygen Demand), NH₄⁺-N, TN and TP were 35.3 ± 4.8 mg/L, 0.4 ± 0.1 mg/L, 1.3 ± 0.2 mg/L and 0.14 ± 0.04 mg/L, respectively, which met the class V surface water standards in China. This result further supports the feasibility of this approach for water quality treatment. Moreover, placing ecological treatment facilities on roofs can help alleviate the problems associated with excessive land use and urban heat island effect. These systems have the advantages of low operating costs, conservation of building energy and reduction in emissions.

4. Conclusions

The novel PBT-GR system proposed in the current study shows promise for the collection, treatment and reuse of road rainwater runoff. Consequently, we can reach the following conclusions:

(1)

The hydraulic load of the green roof should be 0.3 m³·(m²·d)⁻¹ to maximize the utilization of nitrogen and phosphorus resources. Under a typical rainfall, the initial interception of road rainfall–runoff by the porous pavement and the secondary purification by the bioretention system can significantly improve water quality. The removal rates of SS, N, P, Pb and Zn by the porous pavement of 63.9%, 18.8%, 27.8%, 36.8% and 30.4%, respectively, improved to 90.5%, 53.3%, 73.8%, 80.4% and 72.1%, respectively, after treatment by the bioretention system.

(2)

Removal of N, P and heavy metals was significantly higher in the bioretention unit compared to the porous pavement and green roof units (p < 0.05), whereas the porous asphalt pavement achieved significantly higher SS removal compared to the bioretention and green roof units (p < 0.05). The PBT-GR system significantly reduced the pollution level of road rainfall–runoff, with the concentrations of SS, N, P and heavy metals in the effluent of the system as 10.1 mg/L, 1.2 mg/L, 0.11 mg/L, 0.08 mg/L and 0.1 mg/L, respectively, within the class V surface water standards in China. The heavy metal contents of water spinach planting using the effluent of the PBT-GR system met the national standard.

(3)

The estimation of investment required for the PBT-GR system indicated that the investment cost can be recovered within 3 years, thereby indicating the good economic viability of the system. The current study showed that the PBT-GR system was able to effectively treat and utilize road rainfall–runoff and that the system has the advantages of low operation cost, a small volume and density and is environmentally friendly. Therefore, the system is likely to be widely used.

(4)

However, use of the PBT-GR system should consider the source of irrigation rainwater during a drought period, the appropriate selection of vegetables to be planting hydroponically, the specific mechanisms of degradation of pollutants and the physiological and ecological changes experienced by vegetables planting using this system.