Listen

Bioretention cells have been widely used in stormwater management. Media plays an important role in stormwater runoff volume reduction and pollutant removal. A novel bioretention media of rock wool has been proposed, and its effect on stormwater runoff volume reduction was investigated via a laboratory-scale experiment. The results showed that compared with the conventional bioretention (CB) using medium sand as media, the volume capture ratio of annual rainfall (VCRAR) of the rock wool bioretention (RWB) was 2.66–5.5% higher, and the peak flow reduction rate was 2.97–11.32% higher. The removal efficiency of the RWB on COD, NH4+ -N, TN, and TP were 25–30%, 10–31%, 20–40%, and 5–14% higher than the CB, respectively. The removal efficiency of the RWB on both Pb and Zn was >99%. The microbial high-throughput sequencing results showed that the RWB can provide a better breeding environment for Bacteroides and Actinobacteria, which is conducive for the removal of pollutants. The RWB had a better stormwater runoff volume reduction rate and pollutant removal efficiency than the CB, therefore, it could be used as a better media for bioretention cells to improve stormwater management efficiency by using rock wool, obtained from construction waste, as the media.

Listen

  • Rock wool has large porosity, superior permeability, and better water retention performance.

  • Rock wool filler has a higher runoff volume reduction rate than the medium sand.

  • Rock wool bioretention has excellent and stable runoff pollutant removal efficiency.

  • The presence of Bacteroides and Actinobacteria in the rock wool was higher than in the medium sand.

peak flow rate, runoff pollutant removal, sponge city, stormwater runoff, volume capture ratio of annual rainfall
Listen

With rapid urbanization, impervious areas have significantly increased in cities, resulting in an increase in stormwater runoff peak flow rate and total discharge volume (Davis Allen 2008). Concurrently, urban flooding and stormwater runoff pollution are becoming increasingly serious throughout the world. For this, China has proposed sponge city construction, advocating for priority use of natural measures for stormwater runoff retention, purification, and harvesting. Among them, bioretention is one of a widely used measures in sponge city construction for retention and purifying stormwater runoff through the synergistic use of plants, soil, media. and microorganisms. Purified stormwater can not only supplement groundwater but also be reused for green belt irrigation or discharged into the municipal drainage pipe network (Li & Davis 2009). Lots of research has found that the characteristics of bioretention media is crucial for the pollutant removal and runoff detention effect (Barrett et al. 2013). Such as, more than 90% of heavy metals in stormwater runoff are removed in the media layer (Muthanna et al. 2007), and some studies have shown that different media have specific removal abilities for different pollutants in stormwater runoff. For example, Dietz & Clausen (2006) found that bioretention with sand as media can remove ammonia nitrogen and total nitrogen (TN) in high efficiency, but the removal efficiency of total phosphorus (TP) was low and unstable, and even increased in the effluent compared with influent.

In order to improve the pollutant removal efficiency of bioretention facilities, there have been numerous studies exploring the efficacy of different types of bioretention media. For instance, Randall & Bradford (2013) examined and compared the adsorption capacity of different bioretention media, including sandy soil mix, soil amended with alum-based drinking water treatment residuals, oxide-coated media and lanthanum-modified bentonite, the results showed that sandy soil mix had higher TP and TN adsorption capacity with a removal rate of about 75.5 and 53.4%. Additionally, vermiculite had a higher COD, TN, and TP removal efficiency, with removal efficiency of 60–70, 80–90, and 60–70%, respectively. Davis et al. (2001) research showed that the bioretention cell consists of porous soil, a topping layer of hardwood mulch and plant species, can achieve high pollutants removal efficiency for infiltration, adsorb and adsorption, and biodegradation performance. The removal efficiency of TKN, and TP all ranged from 60 to 80%. Apart from single media, some studies have investigated the characteristics of composite bioretention media. For instance, Pan et al. (2020) studies found that a combination of sawdust compost, zeolite, and anthracite in specific proportions as composite media of bioretention cell could significantly enhance COD and TP removal efficiency in stormwater runoff, ranging from 70.5 to 87.5 and 50.0 to 79.3%, respectively. So, modified media have better pollutant removal efficiency than original sole media.

Another goal for the bioretention cell design was to capture the runoff volume and alleviate the local flooding, The types of bioretention media have an important effect on the runoff volume reduction rate (Winston et al. 2016). The infiltration rate of the media and the internal water storage volume are the sensitive parameters affecting the runoff volume reduction rate of bioretention cells (Winston et al. 2016; Tirpak et al. 2019). However, some research was carried out to recognize the bioretention hydraulic characteristics, especially permeability with time elapsed, runoff detention capacity, and runoff holding capacity, as well as the coupled correlation between runoff volume capture and pollutant removal efficiency. Additionally, in the engineering application of bioretention, commonly used bioretention media generally has well adsorbability and large porosity, but poor retention capacity (Siriwardene et al. 2007; Lim et al. 2015), this leads to an increase in irrigation water consumption for bioretention cells, or affect the normal growth of plants. Therefore, it is important to find a new medium with high pollutant removal efficiency, permeability, and retention capacity.

Rock wool is a widely used building exterior wall thermal insulation and fire protection material in China, with annual production of approximately 3.8 million tons by 2023 (Yan et al. 2021). It is made from high-quality basalt, white jade and other materials, which are melted above 1,450 °C and then centrifuged into fibers at high speed, and sprayed with a quantitative caking agent, dust oil and other materials, through the collection, pressing, solidification, and cutting processes. After adding hydrophilic additives, eco-porous fiber is formed and has excellent permeability, absorption and moisture-holding characteristics, and porosity of > 95% (Yan et al. 2021). Compared with sand, rock wool has higher porosity, adsorption ability, and water retention capacity. Additionally, rock wool is beneficial to plant growth because of its high water retention capacity and air permeability when used as a substrate for plant growth (Wu et al. 2014).

Rock wool by construction waste was selected as a bioretention media, and the main goals of this work included (1) to investigate the runoff volume reduction rate of the rock wool as bioretention media, (2) to investigate runoff pollutants removal efficiency of the rock wool as bioretention media, (3) to compare the runoff volume reduction rate and runoff pollutants removal efficiency between the rock wool and medium sand as bioretention media. So as to provide valuable insights into the application of rock wool as a bioretention media.

Listen

Experimental device

Listen
Two polyvinyl chloride (PVC) bioretention cells were designed, one with conventional bioretention (CB, Figure 1(a)) and the other with rock wool bioretention (RWB; Figure 1(b)). Both were 80 cm tall and with a cross-section of 40 × 40 cm. In each, the overflow outlet was placed 20 cm above the planting soil layer. From top to bottom, the CB device consisted of a detention layer (200 mm), a planting soil layer (200 mm, loess soil), a media layer (300 mm, medium sand), a drainage layer (50 mm, gravel, Φ4–6 mm), and permeable geotextile (250 g/m2) was set between the media and the adjacent structural layer. Meanwhile, from top to bottom the RWB comprised of a detention layer (200 mm), a planting soil layer (200 mm, loess soil), a media layer (300 mm, rock wool), and a drainage layer f(50 mm gravel, Φ4–6 mm). Iris lactea Pall., known for its excellent drought resistance, strong soil fixation and water storage abilities, air humidity regulation, and pollutant removal properties, was planted in the planting soil layer in both experimental devices.
Figure 1
Bioretention devices: (a) conventional bioretention (CB) and (b) rock wool bioretention (RWB).
View largeDownload slide

Bioretention devices: (a) conventional bioretention (CB) and (b) rock wool bioretention (RWB).

Figure 1
Bioretention devices: (a) conventional bioretention (CB) and (b) rock wool bioretention (RWB).
View largeDownload slide

Bioretention devices: (a) conventional bioretention (CB) and (b) rock wool bioretention (RWB).

Close modal
The rock wool adopted in the experiment is shown in Figure 2(a). It is composed of an interlocked filamentous structure matter with higher void space (Figure 2(b)), the surface of which is rough, with numerous papillary protrusions that increase its surface area (Figure 2(c)), making it receptive to pollutant attachment. The density, compressive strength and porosity of rock wool were 121.43 kg·m−3, 3,000 kg·m−2, 97.65%, respectively. Those for medium sand were 1,470 kg·m−3, 3,300 kg·m−2, 46%, respectively.

  • (1) Permeability coefficients

Figure 2
Scanning electron microscope rock wool images: (a) rock wool; (b) rock wool fiber network structure; and (c) rock wool fiber surface structure.
View largeDownload slide

Scanning electron microscope rock wool images: (a) rock wool; (b) rock wool fiber network structure; and (c) rock wool fiber surface structure.

Figure 2
Scanning electron microscope rock wool images: (a) rock wool; (b) rock wool fiber network structure; and (c) rock wool fiber surface structure.
View largeDownload slide

Scanning electron microscope rock wool images: (a) rock wool; (b) rock wool fiber network structure; and (c) rock wool fiber surface structure.

Close modal
The constant head method was adopted to calculate the permeability coefficient of rock wool and medium sand according to formula (1).
(1)
where k is the permeability coefficient, Q is the volume flow through the rock wool, A is the device cross-sectional area, L is rock wool thickness, △h is the water head difference.

The permeability coefficient of different sides of rock wool is different due to the degree of compaction during production. The specific permeability coefficient from different directions of rock wool are shown in Table 1. It can be found that the permeability coefficient from side 3 direction influent was minimum with the average value of 1.36 mm/s, and has a high water retention capacity, so the runoff infiltration along side 3 direction in the RWB. The infiltration rate of medium sand was 2.0 × 10−2 mm/s.

  • (2) Water retention and release rate

Table 1

Rock wool permeability coefficients

Group 1 (mm/s)Group 2 (mm/s)Group 3 (mm/s)
Side 1 2.76 2.70 2.74 
Side 2 2.24 2.17 2.22 
Side 3 1.37 1.36 1.36 
Group 1 (mm/s)Group 2 (mm/s)Group 3 (mm/s)
Side 1 2.76 2.70 2.74 
Side 2 2.24 2.17 2.22 
Side 3 1.37 1.36 1.36 
View Large

The rock wool block with a volume of L × B × H = 0.5m × 0.5m × 0.5 m was put into a device with a size of 0.5m × 0.5m × 0.65 m, and the water was fed until the water surface just immerged the surface of the rock wool. After that, the bottom valve was opened for drainage, and the drainage volume with time elapse was recorded. Its maximum water storage volume was 112.88 L, water storage rate was 90.3%, and water retention rate was 47.24%. The water release rate of the rock wool gradually decreased with time elapse, and the water release volume was 59.55 L, accounting for 52.46% of storage volume (Figure 3). The water retention rate and water released rate of medium sand was 19 and 81%, respectively.
Figure 3
Rock wool block water release process after full storage.
View largeDownload slide

Rock wool block water release process after full storage.

Figure 3
Rock wool block water release process after full storage.
View largeDownload slide

Rock wool block water release process after full storage.

Close modal

Experimental scheme

Listen

In accordance with the Technical Guide for Sponge City Construction (China 2015), the optimal ratio of the bioretention cell's area to the catchment area was 1:10–1:20, and 1:10 was adopted in the experiment. For the bioretention cell most often used in road green belts, the assumption is that the land type of the catchment was an urban road, so the runoff coefficient adopts 0.8. To simulate the catchment runoff process under different return periods, rainfall duration and rainfall peak coefficients, the experimental device inflow rate was adjusted by regulating the peristaltic pump speed (Rever YZ-35).

  • (1) Runoff volume reduction experiment

For rainfall event simulation, the Beijing rainfall intensity formula and Chicago rain pattern were selected, and their return period, rainfall duration and rainfall peak coefficients were adopted as experimental variables (Table 2).

Table 2

Rainfall parameters adopted in experiment

FactorsRainfall duration (h)Rainfall peak coefficientReturn period (years)
Return period 2 0.4 1 
2 0.4 3 
2 0.4 5 
2 0.4 10 
Rainfall duration 1 0.4 5 
3 0.4 5 
5 0.4 5 
12 0.4 5 
Rainfall peak coefficient 2 0.3 5 
2 0.5 5 
2 0.7 5 
FactorsRainfall duration (h)Rainfall peak coefficientReturn period (years)
Return period 2 0.4 1 
2 0.4 3 
2 0.4 5 
2 0.4 10 
Rainfall duration 1 0.4 5 
3 0.4 5 
5 0.4 5 
12 0.4 5 
Rainfall peak coefficient 2 0.3 5 
2 0.5 5 
2 0.7 5 
View Large

The rainfall process under different experimental conditions is shown in Figure 4. The rainfall intensity in Beijing was calculated according to formula (2) and formula (3), and the application conditions of Equation (2) is for t ≤ 120 min, P ≤ 10 a, and Equation (3) is for t > 120 min, P ≤ 10 a.
(2)
(3)
where q is the design rainfall intensity, L·(s·hm2)−1; P is the design rainfall return period; t is the rainfall duration, min.

  • (2) Pollutants in simulated stormwater runoff

Figure 4
Rainfall hydrograph under different rainfall conditions: (a) return period; (b) rainfall duration; and (c) rainfall peak coefficient.
View largeDownload slide

Rainfall hydrograph under different rainfall conditions: (a) return period; (b) rainfall duration; and (c) rainfall peak coefficient.

Figure 4
Rainfall hydrograph under different rainfall conditions: (a) return period; (b) rainfall duration; and (c) rainfall peak coefficient.
View largeDownload slide

Rainfall hydrograph under different rainfall conditions: (a) return period; (b) rainfall duration; and (c) rainfall peak coefficient.

Close modal

Pollutant concentrations in simulated stormwater runoff were in accordance with the monitoring data in the urban main road in Beijing (Wang et al. 2019). KH2PO4 (AR), NH4Cl (AR), KNO3 (AR), and C6H12O6 (AR) were adopted as phosphorus, nitrogen, and carbon sources in the simulated stormwater runoff. Low pollutant concentration (LC), medium pollutant concentration (MC), and high pollutant concentration (HC) were set to evaluate the pollutant removal efficiency of different bioretention media, as shown in Table 3. Rainfall events with a rainfall duration of 1 h and a return period of 5a were used.

  • (3) Pollutants monitoring methods

Table 3

Stormwater runoff pollutant concentrations

Pollutants concentrationTP (mg·L−1) (mg·L−1)TN (mg·L−1)COD (mg·L−1)Pb (mg·L−1)Zn (mg·L−1)
LC 1.0 ± 0.1 3.0 ± 0.2 11.0 ± 0.2 100 ± 10 0.1 ± 0.01 0.2 ± 0.01 
MC 2.0 ± 0.1 6.0 ± 0.2 22.0 ± 0.2 200 ± 10 0.2 ± 0.01 0.4 ± 0.01 
HC 3.0 ± 0.1 9.0 ± 0.2 33.0 ± 0.2 300 ± 10 0.3 ± 0.01 0.6 ± 0.01 
Pollutants concentrationTP (mg·L−1) (mg·L−1)TN (mg·L−1)COD (mg·L−1)Pb (mg·L−1)Zn (mg·L−1)
LC 1.0 ± 0.1 3.0 ± 0.2 11.0 ± 0.2 100 ± 10 0.1 ± 0.01 0.2 ± 0.01 
MC 2.0 ± 0.1 6.0 ± 0.2 22.0 ± 0.2 200 ± 10 0.2 ± 0.01 0.4 ± 0.01 
HC 3.0 ± 0.1 9.0 ± 0.2 33.0 ± 0.2 300 ± 10 0.3 ± 0.01 0.6 ± 0.01 
View Large

Monitoring methods of COD, ⁠, TN, TP, Zn, and Pb are shown in Table 4.

  • (4) Microbiological community analysis

Table 4

Pollutants monitoring methods

PollutantsMonitoring methods
COD Potassium dichromate method (GB 11914-89) 
 Nessler's reagent spectrophotometry (HJ 535-2009) 
TN Alkaline potassium persulfate digestion ultraviolet spectrophotometry (HJ 636-2012) 
TP Ammonium molybdate spectrophotometry (GB 11893-89) 
Zn, Pb Atomic absorption spectrophotometry (GB 7475-87) 
PollutantsMonitoring methods
COD Potassium dichromate method (GB 11914-89) 
 Nessler's reagent spectrophotometry (HJ 535-2009) 
TN Alkaline potassium persulfate digestion ultraviolet spectrophotometry (HJ 636-2012) 
TP Ammonium molybdate spectrophotometry (GB 11893-89) 
Zn, Pb Atomic absorption spectrophotometry (GB 7475-87) 
View Large

After the experiment was completed, the medium of the two bioretention facilities was sampled by three-point sampling method. The samples were mixed evenly and stored in dry ice, and the Microbiological community was analyzed by High-throughput sequencing. The analysis process was as follows : (1) After the genomic DNA extraction was completed, the extracted genomic DNA was detected by 1% agarose gel electrophoresis ; (2) PCR amplification, PCR using TransGen AP221-02: TransStart Fastpfu DNA Polymerase; the instrument was ABI GeneAmp® 9,700. All samples were carried out according to the formal experimental conditions, and each sample had three replicates. The PCR products of the same sample were mixed and detected by 2% agarose gel electrophoresis. The PCR products were recovered by AxyPrep DNA gel recovery kit and eluted with Tris _ HCl. 2% agarose was used for electrophoresis detection. (3) Fluorescence quantification was performed. According to the preliminary quantitative results of electrophoresis, the PCR products were detected and quantified by QuantiFluor TM-ST blue fluorescence quantitative system, and then the corresponding proportions were mixed according to the sequencing requirements of each sample. (4) The Illumina library was constructed with TruSeqTM DNA Sample Prep Kit. (5) Final Illumina sequencing.

Listen

Runoff volume reduction rate under different rainfall parameters

Listen
The volume capture ratio of annual rainfall (VCRAR) (fV) and peak reduction rate (Rpek) are calculated according to formula (4) and formula (5).
(4)
where is the total volume of rainwater discharged from the bioretention facility, L; is the total volume of rainwater entering the bioretention facility, L.
(5)
where is the peak flow rate of outflow, L·s−1; is the peak flow rate of inflow, L·s−1.

  • (1) Rainfall return periods

The runoff hydrograph under different return periods of the two bioretention cells is shown in Figure 5. It can be found that when the rainfall return period was 1 year, there was no overflow in the RWB. However, when rainfall return periods increased to 3, 5, and 10 years, overflow occurred at an earlier rainfall stage and overflow volume gradually increased with return periods increasing. At the same time, the peak time was advanced at about 55–60 min. When rainfall return periods were 1, 3, 5, and 10 years, the accumulated infiltration volume of CB was 42.489, 46.121, 44.832, and 45.781 L, and for the RWB was 43.183, 50.879, 49.337, and 51.731 L, respectively. So, the RWB exhibited higher infiltration performance compared to the CB, and its overflow volume and peak flow rate were all lower than that in the CB. The reason is that rock wool possesses superior water retention capabilities than medium sand and has a higher porosity.
Figure 5
Runoff hydrograph under different return periods: (a) 1a; (b) 3a; (c) 5a; and (d) 10a.
View largeDownload slide

Runoff hydrograph under different return periods: (a) 1a; (b) 3a; (c) 5a; and (d) 10a.

Figure 5
Runoff hydrograph under different return periods: (a) 1a; (b) 3a; (c) 5a; and (d) 10a.
View largeDownload slide

Runoff hydrograph under different return periods: (a) 1a; (b) 3a; (c) 5a; and (d) 10a.

Close modal
The VCRAR and peak flow reduction rate of the two bioretention cells was investigated at different return periods, the results are shown in Figure 6. It can be found that when rainfall return periods were 1, 3, 5, and 10 years for CB, the VCRAR was 97.41, 67.22, 56.54, and 56.28%, respectively. Additionally, the peak flow reduction rate was 89.83, 38.59, 37.86, and 37.75%, respectively. The results of Jiang et al. (2017) showed that the VCRAR of traditional bioretention facilities was 54.08–98.25%. Meanwhile, the VCRAR of the RWB was 100, 70.34, 59.65, and 58.75%, respectively. Its corresponding peak flow reduction rates were 100, 39.90, 39.01, and 38.87%, respectively. Notably, the rock wool permeability coefficient (1.36 × 10−1 cm·s−1) is significantly higher than that of medium sand (2.00 × 10−3 cm·s−1). However, there was little difference between CB and RWB infiltration and overflow rates. The above phenomenon can be mainly attributed to the close time interval of rainfall during the experiment, the short rainfall interval and the high water content of the two bioretention cells (Gao et al. 2018). Consequently, the potential benefits offered by the high water retention capacity of rock wool media were underutilized.

  • (2) Rainfall duration

Figure 6
Runoff volume reduction rate under different return periods: (a) peak flow reduction rate and (b) VCRAR.
View largeDownload slide

Runoff volume reduction rate under different return periods: (a) peak flow reduction rate and (b) VCRAR.

Figure 6
Runoff volume reduction rate under different return periods: (a) peak flow reduction rate and (b) VCRAR.
View largeDownload slide

Runoff volume reduction rate under different return periods: (a) peak flow reduction rate and (b) VCRAR.

Close modal
RWB and CB infiltration outflow time varies with different rainfall durations were shown in Figure 7. It can be found that when rainfall duration was 1, 3, 5, and 12 h, RWB infiltration outflow times were 11, 21, 17, and 15 min, respectively. For that compared to RWB, CB had longer infiltration outflow times of 28, 40, 42, and 44 min for above different rainfall durations, respectively. From the above results, RWB infiltration outflow time was earlier than CB, which can be attributed to the higher permeability coefficient of rock wool.
Figure 7
Runoff hydrographs under different rainfall durations: (a) 1 h; (b) 3 h; (c) 5 h; and (d) 12 h.
View largeDownload slide

Runoff hydrographs under different rainfall durations: (a) 1 h; (b) 3 h; (c) 5 h; and (d) 12 h.

Figure 7
Runoff hydrographs under different rainfall durations: (a) 1 h; (b) 3 h; (c) 5 h; and (d) 12 h.
View largeDownload slide

Runoff hydrographs under different rainfall durations: (a) 1 h; (b) 3 h; (c) 5 h; and (d) 12 h.

Close modal
The VCRAR and peak flow reduction rate of the two different bioretention cells were investigated at different rainfall durations, the results are shown in Figure 8. It can be found that when rainfall duration was 1, 3, 5, and 12 h, the VCRAR of CB was 50.67, 65.63, 66.67, and 74.88%, respectively. Additionally, peak flow reduction rates were 15.49, 49.87, 8.88, and 33.17%, respectively. The VCRAR of RWB was 55.33, 68.75, 69.49, and 78.21%, respectively. Additionally, peak flow reduction rates were 31.83, 51.51, 18.36, and 34.21%, respectively. So, the VCRAR of both bioretention cells increased with increasing rainfall duration. However, the peak flow reduction rate did not show obvious regularity under different rainfall durations. The bioretention cell filling layer was not saturated before the runoff peak time. However, when the rainfall duration was 1, 3, 5, and 12 h, the CB infiltration rate gradually increased and then tended to be stable at 50, 75, 140, and 305 min, respectively, and that of RWB at 45, 65, 130, and 295 min, respectively. When the filling layer is saturated with water, bioretention reaches the steady infiltration stage, therefore, the overflow rate will increase as the inflow rate increases (Kim et al. 2019). So, the proportion of overflow volume to total inflow volume decreased with increasing rainfall duration. Therefore, VCRAR increased with increasing rainfall duration.

  • (3) Rainfall peak coefficients

Figure 8
Runoff volume reduction rate under different rainfall durations: (a) peak flow reduction rate and (b) VCRAR.
View largeDownload slide

Runoff volume reduction rate under different rainfall durations: (a) peak flow reduction rate and (b) VCRAR.

Figure 8
Runoff volume reduction rate under different rainfall durations: (a) peak flow reduction rate and (b) VCRAR.
View largeDownload slide

Runoff volume reduction rate under different rainfall durations: (a) peak flow reduction rate and (b) VCRAR.

Close modal
RWB and CB runoff hydrographs vary with different rainfall peak coefficients shown in Figure 9. It can be found that when rainfall peak coefficients were 0.3, 0.5, and 0.7, RWB peak overflow rates were 49.2, 60.2, and 73.0 mL·s−1, respectively. So, as the rainfall peak coefficient increased, the RWB peak overflow rate also increased. Specifically, for rainfall peak coefficients of 0.3, 0.5, and 0.7, the peak overflow rate of CB was 51.0, 61.1, and 73.8 mL·s−1, respectively. So, the peak overflow rate of CB was slightly larger than that for RWB under different rainfall peak coefficients.
Figure 9
Runoff hydrograph under different rainfall peak coefficients: (a) 0.3; (b) 0.5; and (c) 0.7.
View largeDownload slide

Runoff hydrograph under different rainfall peak coefficients: (a) 0.3; (b) 0.5; and (c) 0.7.

Figure 9
Runoff hydrograph under different rainfall peak coefficients: (a) 0.3; (b) 0.5; and (c) 0.7.
View largeDownload slide

Runoff hydrograph under different rainfall peak coefficients: (a) 0.3; (b) 0.5; and (c) 0.7.

Close modal
The VCRAR and peak flow reduction rate of the two bioretention cells were investigated at different rainfall peak coefficients, the results are shown in Figure 10. It can be found that When the rainfall peak coefficient increased from 0.3 to 0.7, the VCRAR of CB decreased from 62.95 to 55.69% and that for RWB decreased from 65.73 to 59.02%. A similar trend was observed for peak flow reduction rates, when the rainfall peak coefficients were 0.3, 0.5, and 0.7, that for CB decreases were 52.93, 6.79, and 6.14%, and for RWB decreases were 53.61, 7.71, and 6.94% respectively. With increasing rainfall peak coefficient, the overflow volume of the two bioretention cells gradually increased, while the VCRAR and peak flow reduction rate decreased. RWB runoff volume reduction rates were all greater than that of CB under different rainfall peak coefficients. As the rainfall peak coefficient increased, peak time was delayed, and the ratio of the overflow volume to the total inflow volume increased, by 23.64, 47.16, and 54.79%, when rainfall peak coefficients were 0.3, 0.5, and 0.7, respectively. Therefore, the water content of the media was higher when peak time was delayed, and the media infiltration rate was decreased. So, when the peak coefficients increased, the overflow volume also increased.
Figure 10
Runoff volume reduction rate under different rainfall peak coefficients: (a) peak flow reduction rate and (b) VCRAR.
View largeDownload slide

Runoff volume reduction rate under different rainfall peak coefficients: (a) peak flow reduction rate and (b) VCRAR.

Figure 10
Runoff volume reduction rate under different rainfall peak coefficients: (a) peak flow reduction rate and (b) VCRAR.
View largeDownload slide

Runoff volume reduction rate under different rainfall peak coefficients: (a) peak flow reduction rate and (b) VCRAR.

Close modal

Removal efficiency of typical pollutants in stormwater runoff

Listen

  • (1) COD

The COD removal efficiency by CB and RWB under different pollution levels was shown in Figure 11, it can be found that when the influent COD concentration in LC, MC, and HC levels, which for CB was73.92, 85.74, and 61.82%, respectively. Meanwhile, that for RWB were 88.91, 92.87, and 93.04%, respectively. So, COD removal efficiency by RWB was higher than CB under different influent COD concentrations. The above CB experimental results were in agreement with the literature, such as, Hunt et al. (2006) monitored a raingarden in Beijing, where COD removal efficiency was ranged in 35–96.2%. Other studies showed that average COD removal efficiency by CB ranged from 59–80% (Davis et al. 2001; Qiu et al. 2019). When particular modified media were used, COD removal ranged from 85–89% (Barrett et al. 2013). Through a comparison of COD removal ranges between RWB and literature, it can be found that COD removal efficiency by RWB was 25–30% higher, and more stable than other media, especially under high inflow COD concentrations.

  • (2) Nitrogen

Figure 11
COD removal efficiency by CB and RWB under different pollution levels.
View largeDownload slide

COD removal efficiency by CB and RWB under different pollution levels.

Figure 11
COD removal efficiency by CB and RWB under different pollution levels.
View largeDownload slide

COD removal efficiency by CB and RWB under different pollution levels.

Close modal
The removal efficiency by CB and RWB under different pollution levels was shown in Figure 12(a), it can be found that as influent concentration increased from 3 to 9 mg·L−1, the average CB removal rate of gradually increased, from 68.36 to 86.03%, and that by RWB increased from 89.78 to 95.31%. With increasing concentration in influent, the effluent RWB concentration changed little, from 0.1 to 0.8 mg·L−1. The above CB experimental results were in agreement with the literature. Such as, Li & Davis (2014) research found that the influent concentration only slightly affected the bioretention effluent concentration. For in stormwater runoff was positively charged, while rock wool was pressed from basalt fiber, which was generally negatively charged. Therefore, adopting rock wool as bioretention media can improve removal efficiency through electrostatic interaction.
Figure 12
and TN removal efficiency by CB and RWB under different pollution levels (a) and (b) TN.
View largeDownload slide

and TN removal efficiency by CB and RWB under different pollution levels (a) and (b) TN.

Figure 12
and TN removal efficiency by CB and RWB under different pollution levels (a) and (b) TN.
View largeDownload slide

and TN removal efficiency by CB and RWB under different pollution levels (a) and (b) TN.

Close modal

The TN removal efficiency by CB and RWB under different pollution levels was shown in Figure 12(b), it can be found that when the TN inflow concentration was LC and MC for CB, the removal efficiency of which was ranged in 37–71% and 55–72%. Yin (2016) research also found that the TIN removal efficiency was poor when influent TN concentration was in low level. The average TN removal efficiency by CB was ranged in 50.4–63.39%, while that for RWB was 72.55–80.41% under different inflow TN concentrations, which was 20–40% higher than CB. So, and TN removal efficiency via bioretention greatly based on the media types, some researchers found that and TN removal efficiency by CB was 60–80% and 30–77% (Hunt et al. 2008; Li & Davis 2009). Another researcher found that the removal efficiency of was 80–93%, and TN was 59–80% when zeolite and other mixed media were used (Barrett et al. 2013; Palmer et al. 2013). Both the and TN removal efficiency via rock wool were higher than that of via other media. Moreover, and TN removal efficiency was more stable than other medias, especially under high inflow and TN concentrations.

  • (3) Total phosphorus (TP)

TP removal efficiency by RWB and CB under different pollution levels was shown in Figure 13, it can be found that TP removal efficiency by CB was 74.59–85.87%, and by RWB was 88.38–90.56%. So, TP removal efficiency by RWB was 5–14% higher than that of CB. Hence, bioretention media had an important impact on TP removal, when composite media were used as bioretention media, TP removal efficiency in stormwater runoff was 78.37–93.90% (Zhang et al. 2011). Furthermore, TP removal efficiency of bioretention was stable, and 70–81% of TP could be removed by water treatment residual media (Palmer et al. 2013). While, when biochar and zeolite were used as media, TP removal efficiency ranged from 48 to 73% (Shrestha 2018). Notably, TP removal by rock wool is higher than biochar, sandy, zeolite and other media, due to the fibrous structure and numerous papillary protrusions on the rock wool surface. On the other hand, rock wool is made of basalt fiber, which contains a lot of Al3+ and Fe3+ that can be complex with phosphate to form a precipitate (AlPO4), effectively adhering to the fibrous structure (Poor et al. 2019).

  • (4) Pb and Zn

Figure 13
TP removal efficiency by RWB and CB under different pollution levels.
View largeDownload slide

TP removal efficiency by RWB and CB under different pollution levels.

Figure 13
TP removal efficiency by RWB and CB under different pollution levels.
View largeDownload slide

TP removal efficiency by RWB and CB under different pollution levels.

Close modal
Pb and Zn, as widely existing heavy metals in stormwater runoff, removal efficiency by RWB and CB under different pollution levels were shown in Figure 14. It can be found that Pb removal efficiency by CB was > 99%, and for Zn was 97–98%. Average Pb and Zn removal efficiency by RWB were both > 99%. Zn is more soluble in runoff than Pb because of its lower affinity for Fe, Al and organic matter, which makes it more mobile and difficult to remove than Pb (Li et al. 2021). The above CB experimental results were in agreement with literature, such as Muthanna et al. (2007) research found that bioretention can remove 89–99% both of Pb and Zn in stormwater runoff. Blecken et al. (2009) found that Pb and Zn removal efficiency in bioretention both were 95%. Meanwhile, Pb and Zn average removal efficiency via RWB both were higher than CB and were also more stable, especially under low inflow Zn concentrations (Figure 14(b)). Studies have shown Pb and Zn removal efficiency both about 75–97% by conventional media, such as sandy loam and construction sand (Davis et al. 2001; Glass & Bissouma 2005), and both > 90% by residuals from water supply plant (Qiu et al. 2019). Compared with CB and bioretention with sewage treatment residues as media, RWB has better Pb and Zn removal efficiency.
Figure 14
Pb and Zn removal efficiency by RWB and CB under different pollution levels (a) Pb and (b) Zn.
View largeDownload slide

Pb and Zn removal efficiency by RWB and CB under different pollution levels (a) Pb and (b) Zn.

Figure 14
Pb and Zn removal efficiency by RWB and CB under different pollution levels (a) Pb and (b) Zn.
View largeDownload slide

Pb and Zn removal efficiency by RWB and CB under different pollution levels (a) Pb and (b) Zn.

Close modal

Microbiological community analysis

Listen
The bacterial population structure at the phylum level was obtained by listing the top 6 phyla in relative abundance, and by microbiological community analysis, the results are shown in Figure 15. It can be found that bacterial community structure in rock wool and medium sand was analogous. Relative abundance of Actinobacteria, Bacteroidetes, Cyanobacteria, Firmicutes, Planctomycetes and Proteobacteria accounted for 1.2, 6.9, 3.5, 4.9, 2.1, and 75.7% in CB, and 5.3, 12.2, 0.1, 0.3, 1.7, and 60.4% in RWB, respectively. The total relative abundance of the top 6 phyla in CB and RWB accounted for 94.3 and 79.9%, respectively, and the absolute dominant phylum in the CB and RWB was Proteobacteria, which included a variety of pathogens and nitrogen-fixing bacteria. The relative abundance of those in CB and RWB were 75.7 and 60.4%, respectively, which was similar to the findings of Li et al. (2021). However, the relative abundance of Actinobacteria in CB and RWB was 1.2 and 5.3%, respectively. Some studies have shown that pollutant accumulation has an impact on the growth and metabolic activities of microorganisms, which would promote the growth of Actinobacteria, but inhibit the growth of Proteobacteria (Davis et al. 2001; Zhang et al. 2021). Pollutant accumulation in CB was lower than in RWB, therefore, the relative abundance of Proteobacteria in CB was higher than in RWB, and the relative abundance of Actinobacteria in CB was lower than in RWB. The relative abundance of Bacteroides in CB and RWB was 6.9 and 12.2%, respectively, next only to the Proteobacteria. Bacteroides are important bacteria degrading glucose (C6H12O6), lactose, sucrose and other carbohydrate nutrients. In our experiment, glucose (C6H12O6) was added to the influent, which provided a favorable breeding environment for Bacteroides. Actinobacteria can carry out nitrification (Davis et al. 2001), and absorb nutrients from the plant rhizosphere to enhance plant growth. So, the relative abundance of Bacteroides and Actinobacteria in rock wool was higher, which was beneficial to the removal of COD and nitrogen.
Figure 15
RWB and CB bacterial phylum level community structure (a) CB bacterial community structure and (b) RWB bacterial community structure.
View largeDownload slide

RWB and CB bacterial phylum level community structure (a) CB bacterial community structure and (b) RWB bacterial community structure.

Figure 15
RWB and CB bacterial phylum level community structure (a) CB bacterial community structure and (b) RWB bacterial community structure.
View largeDownload slide

RWB and CB bacterial phylum level community structure (a) CB bacterial community structure and (b) RWB bacterial community structure.

Close modal
Listen

  • (1) Rock wool can be used as an ideal bioretention media for its high density, compressive strength, porosity, infiltration rate, water retention rate and water released rate, which have better runoff volume and peak flow reduction rate than medium sand, improved by 2.66–11.32% under different rainfall conditions.

  • (2) Compared with medium sand bioretention, rock wool bioretention was better at pollutant removal, and improved by 5–40% under different pollution levels. Rock wool was generally negatively charged, so rock wool bioretention efficiently removed ⁠. At the same time, because of the cooperation of Al and TP in rock wool, it also had well TP removal efficiency.

  • (3) Pollutant retention in rock wool made the relative abundance of Proteobacteria lower than in medium sand, and those were 60.4 and 75.7%, respectively. However, the reproduction of Bacteroides and Actinobacteria in rock wool was better, and those were 12.2 and 5.3% higher than in medium sand, respectively, which is beneficial for pollutant removal.

Rock wool used as bioretention media can relieve the conflicts between the high infiltration rate and low retention rate for its high porosity. However, in the future, it will be necessary to further consider the influence of other factors on runoff reduction effects when using rock wool as bioretention media, such as soil layer thickness, plant types, and rainfall interval. The physical characteristics and runoff volume reduction rate and pollutants removal efficiency of rock wool for the long operation of bioretention also need to be recognized in advance.

Listen

The subject does not involve ethical issues.

Listen

All authors agree to participate.

Listen

All authors agree to publish.

Listen

R.Q. wrote the original draft and performed the methodology. J.W. conceptualized the study, performed the methodology, wrote, reviewed, and edited the article. Z.Q. investigated the study and acquired the funds. S.W. wrote, reviewed, and edited the article. T.Y. and J.S. acquired funds and investigated the study. N.T. enhanced the language and edited references.

Listen

This work was supported by the Beijing Municipal Natural Science Foundation (8232022) and Beijing Jinyu Energy Saving and Thermal Insulation Technology (Dachang) Co., Ltd

Listen

The authors declare no competing interests.

Listen

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

Listen

The authors declare there is no conflict.

Barrett
M.
,
Limouzin
M.
 &
Lawler
D.
 (
2013
)
Effects of media and plant selection on biofiltration performance
,
Journal of Environmental Engineering
,
139
,
462
–
470
.
Google Scholar
Crossref
Search ADS
 
Blecken
G.-T.
,
Zinger
Y.
,
Deletić
A.
,
Fletcher
T. D.
 &
Viklander
M.
 (
2009
)
Influence of intermittent wetting and drying conditions on heavy metal removal by stormwater biofilters
,
Water Research
,
43
,
4590
–
4598
.
Google Scholar
Crossref
Search ADS
PubMed
 
China M.o.C.A.o.t.P.s.R.o.
(
2015
)
The Construction Guideline of Sponge City in China-Low Impact Development of Stormwater System (Trail)
.
China Architecture Publishing & Media Co,.Ltd
.
Davis Allen
P.
 (
2008
)
Field performance of bioretention: hydrology impacts
,
Journal of Hydrologic Engineering
,
13
,
90
–
95
.
Google Scholar
Crossref
Search ADS
 
Davis
A.
,
Shokouhian
M.
,
Sharma
H.
 &
Minami
C.
 (
2001
)
Laboratory study of biological retention (Bioretention) for urban storm water management
,
Water Environment Research: A Research Publication of the Water Environment Federation
,
73
,
5
–
14
.
Google Scholar
Crossref
Search ADS
PubMed
 
Dietz
M. E.
 &
Clausen
J. C.
 (
2006
)
Saturation to improve pollutant retention in a rain garden
,
Environmental Science & Technology
,
40
,
1335
–
1340
.
Google Scholar
Crossref
Search ADS
 
Gao
J.
,
Pan
J.
,
Hu
N.
 &
Xie
C.
 (
2018
)
Hydrologic performance of bioretention in an expressway service area
,
Water Science and Technology
,
77
,
wst2018048
.
Google Scholar
 
Glass
C. C.
,
Bissouma
S.
 (
2005
)
Evaluation Of A Parking Lot Bioretention Cell ForRemoval Of Stormwater Pollutants
,
Ecosystems and Sustainable Development V
,
81
,
699
–
708
.
Google Scholar
 
Hunt
W. F.
,
Jarrett
A. R.
,
Smith
J. T.
 &
Sharkey
L. J.
 (
2006
)
Evaluating bioretention hydrology and nutrient removal at three field sites in North Carolina
,
Journal of Irrigation and Drainage Engineering
,
132
,
600
–
608
.
Google Scholar
Crossref
Search ADS
 
Hunt
W.
,
Smith
J.
,
Jadlocki
S.
,
Hathaway
J.
 &
Eubanks
P.
 (
2008
)
Pollutant Removal and Peak Flow Mitigation by a Bioretention Cell in Urban Charlotte, N.C
..
Journal of Environmental Engineering
,
134
,
403
–
408
.
Google Scholar
Crossref
Search ADS
 
Jiang
C.
,
Li
J.
,
Li
H.
,
Li
Y.
 &
Chen
L.
 (
2017
)
Field performance of bioretention systems for runoff quantity regulation and pollutant removal
,
Water, Air, & Soil Pollution
,
228
,
468
.
Google Scholar
Crossref
Search ADS
 
Kim
H.
,
Mallari
K.
,
Baek
J.
,
Pak
G.
,
Choi
H.
 &
Yoon
J.
 (
2019
)
Considering the effect of groundwater on bioretention using the storm water management model
,
Journal of Environmental Management
,
231
,
1270
–
1276
.
Google Scholar
Crossref
Search ADS
PubMed
 
Li
H.
 &
Davis
A. P.
 (
2009
)
Water quality improvement through reductions of pollutant loads using bioretention
,
Journal of Environmental Engineering
,
135
,
567
–
576
.
Google Scholar
Crossref
Search ADS
 
Li
L.
 &
Davis
A. P.
 (
2014
)
Urban Stormwater Runoff Nitrogen Composition and Fate in Bioretention Systems
,
Environmental Science & Technology
,
48
,
3403
-
3410
.
Google Scholar
Crossref
Search ADS
 
Li
Y.-j.
,
Fu
H.
,
Zhang
J.
,
Zhang
Z.
 &
Li
J.-k.
 (
2021
)
Study of Pollutant Accumulation Characteristics and Microbial Community Impact at Three Bioretention Facilities
.
Environmental Science Pollution Research
,
28
,
44389
–
44407
.
Google Scholar
Crossref
Search ADS
PubMed
 
Lim
H. S.
,
Lim
W.
,
Hu
J. Y.
,
Ziegler
A.
 &
Ong
S. L.
 (
2015
)
Comparison of filter media materials for heavy metal removal from urban stormwater runoff using biofiltration systems
,
Journal of Environmental Management
,
147
,
24
–
33
.
Google Scholar
Crossref
Search ADS
PubMed
 
Muthanna
T. M.
,
Viklander
M.
,
Blecken
G.
 &
Thorolfsson
S. T.
 (
2007
)
Snowmelt pollutant removal in bioretention areas
,
Water Research
,
41
,
4061
–
4072
.
Google Scholar
Crossref
Search ADS
PubMed
 
Palmer
E.
,
Poor
C.
,
Hinman
C.
 &
Stark
J.
 (
2013
)
Nitrate and phosphate removal through enhanced bioretention media: Mesocosm study
,
Water Environment Research : A Research Publication of the Water Environment Federation
,
85
,
823
–
832
.
Google Scholar
Crossref
Search ADS
PubMed
 
Pan
J.
,
Liu
Y.
,
Qu
Y. a.
,
Gao
J.
 &
Zhang
X.
 (
2020
)
Removal effect of runoff pollutants by bioretention of composite filler
,
Yellow River (in Chinese)
,
42
,
93
–
99
.
Google Scholar
 
Poor
C.
,
Conkle
K.
,
MacDonald
A.
 &
Duncan
K.
 (
2019
)
Water treatment residuals in bioretention planters to reduce phosphorus levels in stormwater
,
Environmental Engineering Science
,
36
,
265
–
272
.
Google Scholar
Crossref
Search ADS
 
Qiu
F.
,
Zhao
S.
,
Zhao
D.
,
Wang
J.
 &
Fu
K.
 (
2019
)
Enhanced nutrient removal in bioretention systems modified with water treatment residuals and internal water storage zone
.
Environmental Science: Water Research & Technology
,
5
,
993
–
1003
.
Google Scholar
 
Randall
M.
 &
Bradford
A.
 (
2013
)
Bioretention gardens for improved nutrient removal
,
Water Quality Research Journal of Canada
,
48
,
372
.
Google Scholar
Crossref
Search ADS
 
Siriwardene
N. R.
,
Deletic
A.
 &
Fletcher
T. D.
 (
2007
)
Clogging of stormwater gravel infiltration systems and filters: insights from a laboratory study
,
Water Research
,
41
,
1433
–
1440
.
Google Scholar
Crossref
Search ADS
PubMed
 
Tirpak
R. A.
,
Hathaway
J. M.
 &
Franklin
J. A.
 (
2019
)
Investigating the hydrologic and water quality performance of trees in bioretention mesocosms
.
Journal of Hydrology
,
576
,
65
–
71
.
Google Scholar
Crossref
Search ADS
 
Wang
H.
,
Cao
X.
,
Ren
Y.
,
He
Y.
,
Liu
Y.
,
Wang
S.
,
Wang
x.
 &
Li
Z.
 (
2019
)
Analysis on pollution load of rainfall runoff in Beijing urban arterial roads
,
Acta Scientiae Circumstantiae (in Chinese)
,
39
,
1860
–
1867
.
Google Scholar
 
Wu
G.
,
Wang
X.
,
Wu
Z.
,
Dong
Z.
 &
Zhang
G.
 (
2014
)
Durability of basalt fibers and composites in corrosive environments
,
Journal of Composite Materials
,
49
,
873
–
887
.
Google Scholar
Crossref
Search ADS
 
Yin
Y.
 (
2016
)
Impact of Bioretention on Urban Stormwater Runoff
.
Wuhan University of Technology
.
Zhang
Z.
,
Rengel
Z.
,
Liaghati
T.
,
Antoniette
T.
 &
Meney
K.
 (
2011
)
Influence of plant species and submerged zone with carbon addition on nutrient removal in stormwater biofilter
,
Ecological Engineering
,
37
,
1833
–
1841
.
Google Scholar
Crossref
Search ADS
 
Zhang
Z.
,
Li
J.
,
Li
Y.
,
Wang
D.
,
Zhang
J.
 &
Zhao
L.
 (
2021
)
Assessment on the cumulative effect of pollutants and the evolution of micro-Ecosystems in bioretention systems with different media
,
Ecotoxicology and Environmental Safety
,
228
,
112957
.
Google Scholar
Crossref
Search ADS
PubMed
 
© 2024 The Authors
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).