Investigating the influences of concave depths on stormwater runoff and pollution retention of urban grasslands

Over the past few decades, China's urbanization and impervious surfaces have grown rapidly and led to a worsening situation such as urban pluvial flooding and water environment pollution (Jia et al. 2013; Lu et al. 2016), which is joining force to restrict urban sustainable development. However, traditional improved measures on urban stormwater controls (such as increasing the drainage pipe diameters and extending drainage networks to facilitate the rapid runoff discharge) are prohibitively expensive or not effective at mitigating excessive surface runoff during extreme storm events (Pomeroy 2007; Berland et al. 2017). The challenges which urban planners and landscape architects confronted are retrofitting urban landscape for effectively regulating stormwater runoff (De Greef 2005; Ramyar et al. 2021). However, in practice, rainwater management/treatment facilities are usually designed by civil and environmental engineers and do not focus on the basic elements of urban landscape (Backhaus et al. 2012; Ranzato 2017). To improve this situation and promote sustainable stormwater management, the Chinese government announced a ‘Sponge City’ initiative in building urban rainwater infrastructures in 2013 (Che et al. 2015; Zhang et al. 2019). During 2015 and 2016, Chinese government selected 30 cities as pilot sites, and more than 130 cities have formulated ‘Sponge City’ development plans (Chan et al. 2018).

The concept basis and key technology of ‘Sponge City’ are based on green infrastructure (GI) and low impact development (LID) measures (Jia et al. 2017; Li et al. 2019a). The GI/LID strategies emphasized to utilize on site and small-scale source control techniques to maintain and enhance the pre-development hydrologic function of the urban landscape (Benedict & McMahon 2006; Dietz 2007; Baek et al. 2015; Fletcher et al. 2015; Li et al. 2020). The small-scale on-site GI/LID facilities including a green roof, rain barrel/cistern, bioretention system, permeable pavement, concave grassland and wetland channel, etc. (Jiang et al. 2022). Moreover, the ‘Sponge City’ program also aims to transform the urban landscape whilst promoting the greater use of LID facilities that enhance infiltration and storage of urban stormwater such as infiltration trenches, grassed swales, and storage ponds (Xia et al. 2017; Chan et al. 2018; Bai et al. 2019). Especially, the GI concept attempts to influence urban planning and layouts to maximize the benefits of green space (Fletcher et al. 2015).

Green space has an important role in retaining and detaining stormwater runoff (Yao et al. 2015; Zhang et al. 2015; Syrbe & Chang 2018). However, the capacity of green space in reducing direct runoff has not yet effectively employed because of limitations in urban landscape design (Li et al. 2017; Du et al. 2019). Traditionally, the ground surface of urban green space in China is flat or convex and higher than the surrounding roads (Liu et al. 2014). As a result, these flat or convex grasslands are prone to yield surface runoff under the excess infiltration at heavy rainfall events (Du et al. 2019). Concaved grassland (i.e., low elevation or sunken greenbelt), similar to bioretention, rain garden and grassed swale, which with berms deeper than normal lawns was listed as a key measure of the ‘Sponge City’ initiative; it refers to vegetated land that has a lower elevation than its surroundings and can temporarily retain stormwater (Ministry of Housing and Urban-Rural Development 2014). In contrast to a traditional convex design, concave grassland increases rainwater infiltration capacity of grasslands, as well as reduces peak flow and recharges groundwater aquifer (Hood et al. 2007). Thus, concave grassland is a cost-efficient facility of rainwater infiltration, which plays a significant role in rainwater infiltration, storage and utilization (Zhang et al. 2020).

The hydrologic responses and pollution removal performances of concave grasslands differ with various soil properties, concave depths and rainfall characteristics (Wang & Zhang 2015; Winston et al. 2016; Mai et al. 2018; Du et al. 2019). For example, Du et al. (2019) investigated that concave grassland with a depth of 10–20 cm can mitigate direct runoffs by 23.63–98.35%. Cheng et al. (2007) argued that concave grassland with a concave depth of 10 cm could effectively reduce runoff of pluvial floods with 3-year return period. Ye et al. (2001) found that, at a service impervious area ratio of 1:1, runoff reduction rates of a concave grassland for storms of 10, 50, and 100 years were 87.15%, 58.48% and 50.75%, respectively. Liu et al. (2014) showed that runoff volume and peak discharge reduction were respectively achieved of 23.20% and 29.11% by retrofitting grasslands into a concave morphology with a depth of 5 cm. Winston et al. (2016) showed that concave grassland could reduce peak flow by 24%–96% for a 1-year storm event. Mai et al. (2018) investigated that low-elevation greenbelt with 10 cm depth and 5 cm height gullies achieved 43.54%–94.00% runoff volume reduction and 40.59%–97.48% pollutant removal effectiveness. Hunt et al. (2008) showed bioretention effectively attenuated peak flow rates with a mean reduction percentage of 99%. Shrestha et al. (2018) investigated bioretention cells averagely retained suspended solids loads by 94%. Wang & Zhang (2015) found that concave grasslands significantly reduce phosphate, with average total phosphate (TP) reduction rates of 86.28%. Cheng et al. (2009) investigated that the average removal rate of TP from concave grassland outflow was 47.35%.

A series of fundamental and applied questions remain unsolved for the general utilization and development of LID techniques, which actually require further and systematic investigations through extensive practical programs (Vogel et al. 2015; Li et al. 2019b). The main challenge in implementation of concave grasslands is how to efficiently evaluate retention effectiveness, optimize structural configuration and spatial placement (Wang & Banzhaf 2018; Tansar et al. 2022). Although a plethora of studies have been undertaken to assess the runoff reduction effectiveness of LID facilities using pilot experiment and statistical/hydrological model simulation (Li et al. 2019a). However, only few studies were performed through field experiments to investigate the runoff and pollution reduction effectiveness of concave grasslands and explain their runoff and pollution retention mechanism (Elliott & Trowsdale 2007; Ahiablame et al. 2012; Liu et al. 2014). It is difficult to determine the optimal structural configuration and adequate guidelines of concave grasslands without lots of field experiments (Liu et al. 2018). Specifically, the local design criteria and operation conditions have become the critical and urgent questions for the effective design and resilient management of the LID practices (Duan et al. 2016). Therefore, it is necessary to collect practical field data that contribute to finding the most effective strategic solution to overcome barriers for widespread promotion and adoption of LID practices (Shafique & Kim 2015; Eckart et al. 2017).

In the present study, the well-defined scale-based runoff plots of concave grasslands and well-controlled field experimental approaches were conducted in a semi-arid region to investigate the retention performances of runoff quantity and pollution loads of concave grasslands. The impacts of concave depths on stormwater runoff retention and pollution removal rates were quantified, and the minimum area ratios of upslope impervious surface to concave grasslands were estimated. The aim of the present study was: (1) to investigate the effectiveness for runoff volume retention and pollution load reduction of concave grasslands; (2) to analyze the influences of concave depths on runoff and pollution retention of grasslands; (3) to estimate the area ratios of connecting upslope impervious surface to concave grassland. These results are expected to improve our understanding of the hydrological behaviour of concave grasslands, and help urban planners for designing appropriate concave depths and upslope impervious area for retrofitting existing grasslands to be concave.

In this study, a Norton artificial rainfall simulator was set up at 3.5 m above the experimental runoff plots (Figure 1(d)). Veejet 80100 nozzles with 41 kPa water pressure were applied in the spraying systems and spaced 1.1 m apart with a computer that oscillated across the plot to generate a constant rainfall intensity. Specifically, the median volume of raindrop size obtained by this simulator was 2.2 mm, and the uniformity coefficient of rainfall reached more than 0.8. Rainfall depth was monitored by a standard tipping bucket rain gauge (Onset HOBO 0.2 mm Rainfall Smart Sensor, S-RGB), positioned adjacent near and at the same height as that of the experimental runoff plots. In this paper, the runoff outflow from the designated runoff plot was collected in a plastic container below the platform by means of a pipe at its downstream end for continuous monitoring of the weight. In the rainfall experiments, six test repetitions were conducted for each runoff plot. The simulated rainfall events were chosen according to the significant runoff generation process and the capacity of runoff collection by plastic container based on a preliminary experiment (Liu et al. 2019a).

In this study, we used the collected rainwater from a cement ground at the experiment site, and replenished with tap water as the source of supply water for the rainfall simulator. Water quality parameters of concave grassland runoff in this study were focused on total suspended solids (TSS), total nitrogen (TN) and TP concentrations. The outflow from the grassland was collected in a plastic container below the runoff plot through a drainage pipe at its downstream end. Before runoff sampling, the plastic containers and sampling bottles were rinsed three times with distilled water. In each rainfall event, the composite water sample was manually collected after the runoff flow is over. In addition, in order to determine the water quality of input rainwater, six rainwater samples were also collected. The water samples were collected with 0.5 L pre-cleaned polyethylene bottles and immediately stored in a refrigerator at the experimental station. After field experiments, the collected water samples were carried to the laboratory and prepared for testing water quality. The tested methods of water quality were according to the Standard Methods for the Examination of Water and Wastewater, which was published by the American Public Health Association (APHA et al. 2005). TSS concentration was measured using the filtering, drying, and weighing method (Little et al. 2005). TN concentration was measured using the alkaline potassium persulfate digestion and UV spectrophotometric method. TP concentration was measured using the persulfate digestion spectrophotometric method (Liu et al. 2019b).

On the basis of the monitored rainfall and runoff data collected in this study, selected hydrological indicators including time to runoff, runoff discharge depth, mean flow rate, peak flow rate, runoff retention percentage, runoff coefficient and pollutant load reduction rate were calculated and utilized for further analysis. The mean and standard deviation of each runoff indicator and pollutant load reduction rate were quantified for detecting the general features of the runoff retention and related pollution removal characteristics (Liu et al. 2020). The one-way analysis of variance (ANOVA) followed by the Tukey post-hoc test was used to compare the differences in means of time to runoff, runoff retention percentage and peak flow rate for different concave depths and analyzed in the SPSS 17.0. Histograms and line charts were all prepared in the SigmaPlot 14.0.

The reduction effects on TN and TP loads by concave grasslands were attributed to their abilities of runoff volume retention. As demonstrated in Table 3, TSS load reduction rates of concave grasslands were ranged from 50.8% to 97.3%. TN load reduction rate was 49.8% for concave depth of 10 cm. TP load reduction rates were 45.0% and 93.9% for C-5 and C-10 grasslands, respectively. However, TN load reduction rates were ineffective for C-0 and C-5 grasslands, and TP load reduction was ineffective for C-0 grassland. The results were inconsistent with the research presented by Hou et al. (2014), who found the average TP and TN concentrations were reduced by 75.8% and 66.6% by a sunken lawn infiltration system, respectively. As well, Davis et al. (2006) reported good reductions of bioretention in phosphorus (65%–87%) and moderate reductions in nitrogen (49%–59%) concentrations. Fahui et al. (2013) examined the bioretention media performed well with exceptional removal of over 95% of TSS and 82%–96% of phosphorus. Zhang et al. (2020) simulated the water quality effects of rain gardens exhibited the maximum contaminant reduction rates with TSS of 15.5%, TN of 17.3%, and TP of 19.1%. Overall, runoff pollutants load reduction rates of TSS, TN and TP were enhanced along with the increase in concave depths for grasslands. This trend was consistent with the high runoff retention of deeper concave grassland. Consequently, pollutant retention by the concave grasslands was strongly driven by the retention of runoff water volume. The water quality improvement of concave grasslands related with soil media and vegetation species by complex physicochemical reaction, thus their performance will be varied among pollutants types and different structural combination. Therefore, specifically optimizing the growth media, addition of amendments, screening of suitable vegetation, and design alterations should be further researched (Vijayaraghavan et al. 2021).

The present study designed the individual runoff plots of concave grasslands without connecting to surrounding impervious surface inflow. In general, concave grasslands allow stormwater to flow into grasslands from the surrounding roads, pavements and squares, etc. Ideally, the minimum area ratio of upslope impervious surface to concave grassland can be estimated according to maximum runoff retention capacity of flat and concave grasslands. Based on the experimental rainfall-runoff data of grasslands, the calculated maximum runoff retention capacity of C-0, C-5, and C-10 grasslands were 4.2 mm, 20.5 mm, and 79.3 mm, respectively. Table 4 showed the estimated minimum area ratios of upslope impervious surface to grasslands varied from 0.96 to 37.6 under 5, 10, 20 mm rainfall events. The minimum area ratio of upslope impervious surface to grassland C-5 was approximately 1:1 under 20 mm rainfall events. The grassland C-10 could fully control the inflow from a 37.6 times area of upslope impervious surface under rainfall depth of 5 mm. Hou et al. (2020) simulated that when the ponding depth was 15 cm and the flow area ratio was 15:1, runoff control rate of rain garden was 31.9%–100% for 0.5 year to 50 years return periods. Richards et al. (2015) reported that concave grassland with a ratio of 7.5% to catchment area could reduce the flood volume by more than 90%. In this study, the concave grasslands can fully accommodate the assumed inflow runoff from minimum upslope impervious areas without outflow discharge. In application, the estimated minimum impervious area ratios should be enlarged with partial retention goals. These results can help urban planners to design the appropriate criteria of area ratio for upslope impervious surface to promote more efficient concave grassland retrofitting.

The present study experimentally verified runoff volume and pollution load reduction effectiveness of concave grasslands through scale-based runoff plots and simulated storm events, and quantified the impacts of concave depths on stormwater runoff retention and pollution removal rates. Compared with flat grassland, the concave grasslands dramatically delayed runoff generation, retained runoff volume and reduced peak flow rate. Due to runoff volume reduction of concave grasslands depended on their water storage and infiltration capacities, the deeper concave depths led to a delay in runoff generation, and increase the capacity of runoff retention. Results of the regression analysis showed that the concave depths significantly increased time to runoff, and there was also a positive correlation between the concave depths and runoff retention. The developed polynomial equation is applicable to predict runoff retention performance for a concave grassland under similar structural details and rainfall conditions, which help to understand the contribution of concave depth on runoff retention performance. Except TP concentration in grassland with concave depth of 5 cm, mean TSS levels in outflow from concave grasslands were lower as well as TN and TP concentrations were higher than that of input rainwater. Pollution load reduction rates of TSS, TN and TP enhanced along with the increase in concave depths. It indicated that pollutant retention by the concave grasslands was strongly driven by the retention of runoff water quantity. The estimated minimum area ratios of upslope impervious surface to concave grasslands were varied from 0.96 to 37.6 under different rainfall events. In application, the estimated area ratios should be enlarged with partial runoff retention goals.

These experimental results are potentially helpful in better understanding of the mechanism of runoff and pollution retention of concave grasslands, and it would be beneficial for urban planners to raise awareness of retrofitting grasslands to be concave. In stormwater management practices, the concave design is highly proposed in newly built urban grasslands and retrofitting of existing grasslands. Moreover, spatial distribution planning of grasslands should consider draining portion of adjacent impervious surface runoff into the grasslands. These results also assist urban managers and planners for designing an appropriate criteria of concave depth and upslope impervious area ratio to promote more efficient concave grassland retrofitting. As limitations of the results application, results in the present study may largely depend on design structures of concave grasslands, rainfall characteristics, and climatic regions. Future study is needed to experimentally investigate the runoff retention capacity of concave grasslands in different climate regions, continuously simulate the long-term retention performance of concave grasslands, and assess the reduction effectiveness of large-scale implementation of concave grasslands in urban areas for flooding risk control and water quality management.

This study was supported by the National Natural Science Foundation of China [42071051], and the Major Science and Technology Projects of Gansu Province [21ZD4FA008].

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

The authors declare there is no conflict.