Green Roof Systems for Rainwater and Sewage Treatment

Abstract

Green roof systems are regarded as a viable solution for mitigating urban environmental challenges and offering a multitude of environmental benefits. Currently, green roofs are increasingly being utilized for the management of rainwater runoff and wastewater. The integration of decentralized rainwater and sewage on-site treatment technology with urban green buildings is being gradually promoted. Green roofs can also be considered as a form of decentralized rainwater and sewage on-site technology, which holds great potential for widespread adoption in the future. Several studies have suggested that green roofs may serve as a potential source of pollutants; however, there are also studies that clearly demonstrate the efficient removal of nutrients and organic pollutants by green roofs. This article critically examines the existing literature on water treatment aspects associated with green roofs and elucidates their classification and operational mechanisms. Through an analysis of previous research cases, it becomes evident that both substrate and vegetation play a significant role in influencing the treatment performance of green roofs. By designing and configuring appropriate substrate and vegetation, green roofs can play a pivotal role in the purification of water quality. Finally, a brief outlook is presented for the future research directions of green roofs, with the anticipation that green roofs will feature more innovative and environmentally friendly designs, as well as expanded prospects for application.

1. Introduction

The development of green roofs has a rich historical background, tracing its roots back to ancient rooftop gardens. Over 2000 years ago, the initial concept primarily involved the placement of conventional green plants on rooftops. The earliest documented evidence pertaining to green roofs can be found in the Hanging Gardens of Semiramis, located in present-day Syria [1]. In ancient Rome over 2000 years ago, trees were imported for use on local green roofs. Several centuries ago in Scandinavian countries, locals laid grass on rooftops for wind and rain protection, using seaweed as a separator. These early forms of green roofs primarily emphasized aesthetics and practicality in architecture. However, with the economic and social development, developed regions have witnessed a rapid decline in green spaces, resulting in deteriorating environmental conditions. In response to this issue, some countries led by Germany have initiated research on the ecological and environmental aspects of green roofs. In recent years, green roofs have been widely recognized as an effective approach to mitigate urban environmental issues and offer a multitude of ecological benefits [2]. These ecological advantages encompass temperature regulation [3], enhancement of air quality [4], alleviation of the urban heat island effect [5], and promotion of biodiversity [6]. Currently, researchers have conducted several studies on the utilization of green roofs for rainwater runoff management and wastewater treatment. It is anticipated that the future implementation of green roofs for water purification purposes will yield both environmental and economic benefits.

In urban areas, the predominant mode of sewage treatment is centralized water treatment, whereby greywater (wastewater from bathrooms, kitchens, and laundry) and blackwater (wastewater from toilets) are conveyed through sewage pipelines to sewage treatment plants [7]. While this approach is efficient and straightforward, it suffers from low water resource utilization rates as well as high infrastructure construction costs. Furthermore, long-distance transportation of sewage can result in blockages or even leakage that may lead to malodorous water requiring extensive maintenance requirements [8]. Distributed wastewater treatment presents a viable alternative to centralized water treatment, offering enhanced device flexibility and eliminating the need for long-distance water transportation [9]. Moreover, distributed water treatment systems can alleviate the burden on conventional sewage treatment plants [10] as they capitalize on the proximity of wastewater to the treatment system, resulting in reduced pumping requirements and significantly lowered construction and operational costs [11]. Common decentralized water treatment systems encompass constructed wetlands (CWs), aerobic treatment systems, and anaerobic treatment systems. Constructed wetlands are a natural treatment system that employs vegetation, substrate, and biological processes for wastewater treatment. However, the operational feasibility of this system is limited in urban areas with spatial constraints. In anaerobic treatment systems, the lower efficiency of anaerobic bacteria compared to aerobic bacteria results in suboptimal effluent quality attainment, leading to comparatively lower levels of wastewater treatment than other systems [12]. Membrane bioreactor (MBR) technology is widely employed in decentralized water treatment systems. MBR integrates membrane processes with biological wastewater treatment processes to effectively eliminate pollutants from wastewater. Nevertheless, practical implementation of this technology often encounters issues related to membrane fouling, which significantly impairs the efficiency of wastewater treatment processes [13].

In addition to the aforementioned environmental benefits, the implementation of green roof systems can also serve as a decentralized water treatment solution. In comparison to traditional centralized sewage treatment systems, green roof systems offer a more cost-effective and efficient on-site collection and treatment system for rainwater and wastewater. The green roof system can effectively purify effluent water by utilizing vegetation, substrate, and other mechanisms to absorb and filter pollutants. However, several studies have reported that green roofs can potentially act as sources of pollutants, such as nitrogen, phosphorus, and heavy metals [14,15]. Conversely, other research has explicitly highlighted the capacity of green roofs to function as sinks for nitrogen, phosphorus, and certain heavy metals [16,17], thereby mitigating the concentration of pollutants in runoff. According to Chen [18], various studies have indicated that common roofs can contribute to higher levels of runoff pollution compared to green roofs, whereas green roofs can mitigate this pollution to some extent. While some studies have indicated the potential for water pollution caused by green roofs, there is a lack of comprehensive analysis and conclusive findings on this matter. Facing the different results of rainwater and wastewater treatment in green roof systems, this article reviews existing research cases on rainwater runoff management and wastewater purification for reuse. It summarizes the types, operational mechanisms, and designs of current green roof systems. By conducting a thorough analysis and summarization, our aim is to design efficient green roof systems that effectively treat rainwater and sewage for high-quality water purification while minimizing pollution risks. Numerous studies have been conducted on green roof systems for rainwater and sewage treatment, demonstrating that green roof systems have been one of the important techniques for decentralized wastewater treatment. However, the influencing factors of green roof systems are not comprehensively summarized, and discussions on their influencing mechanisms for pollutant removal are still insufficient, which is not beneficial for the development and application of green roof systems. This situation stimulates us to write this comprehensive review article to provide insights and directions for future studies and the design of green roof systems.

2. Methods

To select the most relevant papers on the topic of this review, a comprehensive search was conducted on 15 May 2024, using the keywords ‘green roof’ (or ‘wetland roof’) combined with ‘rainwater treatment’ and ‘sewage treatment’ in the Web of Science (WoS) database. A total of 69,135,14, and 50 papers related to the combined keywords of ‘green roof and rainwater treatment’, ‘green roof and sewage treatment’, ‘wetland roof and rainwater treatment’, and ‘wetland roof and sewage treatment’ were obtained, respectively. By limiting article types to “Article” and “Review”, the number of relevant papers decreased to 68,130,10 and 48, respectively. The emphasis of this review was to summarize the influencing factors of green roof systems on water treatment and discuss the influencing mechanisms. Thus, a detailed analysis was conducted on both the influent and effluent water quality of green roof systems, such as nitrogen, phosphorus, and heavy metals. Based on the literature obtained from the search results, the impact of each factor on both influent and effluent water quality was individually discussed through appropriate comparisons and analyses.

3. Green Roof Systems

3.1. Definition and Classification

The concept of green roofs involves the installation of vegetation on building structures, such as roofs, balconies, or terraces [19]. Similar to conventional green spaces and gardens, these areas are directly exposed to natural elements, including sunlight and rainwater. However, the key distinction lies in the independent soil substrate required for green roofs, which is situated above a designated space rather than being connected to natural soil. In scientific literature, alternative terms used for green roofs include living roofs, eco-roofs, or vegetated roofs [20].

Green roofs often exhibit regional variations in their construction but are typically comprised of five main components: a vegetation layer, a substrate layer, a filter layer, a drainage layer, and a protection layer [21]. The vegetation layer serves as the fundamental element of the green roof and plays a pivotal role in delivering ecological benefits. The substrate layer provides essential nutrients for plant growth while also facilitating rainwater absorption and retention. The filter layer acts as a barrier between the substrate and drainage layers to prevent soil particles from obstructing the drainage system. Serving as both a water storage and drainage system, the drainage layer effectively removes excess water during heavy rainfall while supplying water to plants during dry periods. Lastly, the protection layer safeguards against damage caused by excessive root growth while simultaneously acting as a protective barrier between water and the building’s roof surface. These key components are visually depicted in Figure 1.

Currently, the classification of green roofs primarily relies on substrate thickness, categorizing them into extensive and intensive green roofs. A concise overview of these two roof types is provided based on pertinent studies [23,24]. The two types of green roof patterns are shown in Figure 2.

Extensive green roofs, with a maximum substrate thickness of 15–20 cm, can support limited plant growth due to their shallow depth. They primarily consist of low-growing vegetation such as moss and sedum. Lightweight, porous, and low organic matter substrates are commonly employed for extensive green roofs. These types of roofs can be installed on steep slopes with angles up to 45° and are frequently utilized for fire prevention, insulation, and rainwater management purposes. The thinness, drought resistance, and low load-bearing capacity of the vegetation require no special maintenance at a later stage, with relatively low installation and operation costs.

Intensive green roofs have a substrate thickness of more than 15–20 cm, allowing for a wider range of plant selection. Perennial herbaceous plants, shrubs, or arbors can be chosen for cultivation. The materials utilized in this roofing type are lightweight and possess a low organic concentration. However, due to the increased load resulting from thicker substrates, additional support structures are necessary. Moreover, the installation slope of intensive green roofs is limited and generally remains below 10°. This type of roof is commonly utilized for aesthetic, entertainment, social, and leisure purposes, akin to a rooftop garden. Additionally, it can support a greater variety of organisms and exhibit enhanced biodiversity. However, this particular form necessitates substantial irrigation and maintenance requirements, as well as stringent structural support prerequisites, resulting in elevated construction and operational expenses.

In recent years, researchers have developed innovative green roof systems that differ from typical ones. By integrating horizontal subsurface flow constructed wetlands with green roofs, a shallow wetland roof system has been devised to fully exploit the benefits of constructed wetlands [25,26]. Xu et al. have introduced the hydroponic green roof system, replacing the conventional soil cultivation method with hydroponics [27] and incorporating lightweight fillers to address the issue of heavy substrates in typical green roofs. These shallow bed wetland roof systems and hydroponic green roof systems can collectively be referred to as “wetland roofs”. According to the definition of green roofs, wetland roofs still fall within the conceptual framework of green roofs. Based on their actual construction characteristics and system design, green roofs can be categorized accordingly (Figure 3).

Compared to conventional green roofs, wetland roofs not only facilitate rainwater collection but also offer the potential for wastewater treatment (greywater or blackwater). Wetland roofs share similar advantages with traditional green roofs, as depicted in Figure 4. Therefore, when designing and installing various types of green roofs, it is essential to consider the specific purpose and application in accordance with local climate conditions and the characteristics of the building.

3.2. Operating Mechanism

Green roof systems are recognized as efficient and cost-effective environmental management solutions for rainwater and wastewater treatment, relying on the synergistic effects of plants, substrates, microorganisms, and other factors [29]. The principles underlying water treatment in green roof systems bear resemblance to those of artificial wetlands [30], primarily achieved through physical processes such as sedimentation and filtration, chemical processes including absorption, reaction, and precipitation, as well as biological processes involving microbial activities to accomplish water purification objectives [31].

Plants necessitate a substantial quantity of nutrients throughout their developmental, growth, and reproductive processes. They have the capacity to assimilate nitrogen and phosphorus nutrients via their root system [25]. Moreover, specific plant species possess the capability to remediate water pollutants, particularly metal contaminants, by converting them into non-toxic compounds through intrinsic mechanisms [32]. Additionally, plants can augment pollutant removal by modulating microbial activity in the substrate and modifying its physicochemical composition. The exudation of specific plant root exudates can activate rhizobacteria and enhance pollutant degradation. Well-developed plant roots, along with the biofilm adhering to their surface, often secrete copious amounts of enzymes that expedite the decomposition of water pollutants, thereby accomplishing water purification objectives [33].

The substrate plays a pivotal role in pollutant removal mechanisms, encompassing physical–chemical and biochemical processes. As an integral component supporting plant growth and microbial existence, the substrate exhibits the capacity to absorb and degrade pollutants, thereby exerting a significant influence on the water quality of green roofs [34]. The physical–chemical mechanisms of the substrate involve the initial interception and adsorption of pollutants, followed by potential chemical reactions in certain substrate materials to eliminate corresponding pollutants. With regard to biochemical mechanisms, the substrate provides attachment surfaces for microbial colonization. Microbes form biofilms on these surfaces and employ their metabolic activities to degrade pollutants effectively, consequently enhancing wastewater treatment performance [35].

Microorganisms also play a pivotal role in the remediation of pollutants. Bacteria are essential for nitrogen removal through processes such as assimilation and nitrification–denitrification. The introduction of mycorrhizal fungi into green roof systems not only enhances plant water use efficiency and mitigates drought damage but also reduces nitrogen and phosphorus runoff concentrations. Moreover, they possess the capability to absorb and accumulate heavy metals, thereby enhancing water quality [36]. The specific treatment process of plants, substrates, and microorganisms is summarized in Figure 5.

3.3. Design and Construction

In the construction of green roof systems, careful consideration must be given to the structural load-bearing capacity of the building. Due to limited load standards for roofs, it is necessary to restrict the substrate depth and total weight of green roofs in order to prevent damage caused by excessive weight. In the design and construction process, there is a need for new requirements regarding lightweight materials with low density for green roof substrates in order to minimize weight [38]. During the architectural design process, meticulous planning of drainage systems should be conducted for green roofs to ensure the timely removal of excess water and avoid waterlogging and root-level plant damage. Green roof systems typically consist of drainage layers and protective layers (usually waterproof), which offer some protection but still pose a risk of moisture penetration into the building’s interior. Therefore, when constructing green roofs, emphasis should be placed on ensuring the integrity and long-term feasibility of drainage layers and protective layers. Additionally, sufficient space should be provided for plant root growth in order to prevent structural damage caused by roots [39]. The design and construction of green roofs are influenced by various factors, such as different building structures, original roof slopes, and shapes. Higher slopes often result in increased drainage difficulties, while irregularly shaped roofs add complexity during construction; all these factors must be taken into account during the construction process [40]. However, initial construction costs remain a limiting factor in implementing green roofs despite their potential offset through energy-saving measures over time as well as reduced rainwater management expenses [41].

4. Research Status

4.1. Vegetation

Vegetation constitutes a pivotal element of green roof systems and exerts a significant influence on the water quality of runoff. Owing to their distinct nutrient requirements, utilization efficiency, and impact on nutrient mineralization within the system, different plant types can yield diverse effects on the discharge’s water quality. Due to the unique rooftop environment, characterized by strong winds, intense sunlight, and temperature fluctuations, certain limitations exist for cultivating plants on rooftops. Not all plant species can effectively adapt to such conditions. Moreover, not all plants can endure irrigation with rainwater or wastewater, further narrowing down the available choices for rooftop cultivation. In recent years, the types of plants planted on green roofs are shown in Table 1.

4.1.1. The Influence of Vegetation

In green roof systems, the presence of vegetation exerts a discernible influence on the water quality of the system. Under identical substrate type and thickness conditions, green roofs planted with Ophiopogon japonicus (L. f.) Ker Gawl. exhibit significantly higher effluent TN concentrations compared to unplanted green roofs (p < 0.05) [34]. Furthermore, the TSS concentration of non-planted green roofs (149.11 mg L⁻¹) was found to be three times higher than that of planted green roofs (50.83 mg L⁻¹), and a statistically significant difference between the two was observed (p < 0.01). However, no statistically significant difference was observed in the TP concentration between planted green roofs (0.031 mg L⁻¹) and unplanted counterparts (0.023 mg L⁻¹).

The study conducted by Liao et al. focused on investigating the correlation between plant biomass and the presence of TN and TSS in effluent samples [42]. The findings revealed a positive correlation between plant biomass and TN concentration in the effluent, implying that this association may be attributed to an increased quantity of fallen leaves during the later stages of plant growth. Rapid decomposition of leaf litter and nitrogen mineralization leads to an increase in the TN concentration in effluent. Additionally, it was observed that an increase in plant biomass led to a decrease in the TSS concentration, indicating that larger plants can intercept more runoff, thereby reducing the concentration of suspended solids. In the final measurement, the green roof planted with Agastache foeniculum (Pursh) Kuntze exhibited a significant reduction in TN, dissolved P, dissolved K, dissolved Ca, and dissolved Mg loads by 50%, 28%, 27%, 18%, and 19%, respectively, compared to the control group without plants (p < 0.05).

The study conducted by Park et al. also examined the impact of vegetation on heavy metal concentrations and revealed that green roofs with plant cover exhibited significantly reduced levels of copper, zinc, magnesium, and cadmium in their runoff compared to unplanted green roofs (p < 0.05) [24]. The reduction in heavy metals can be attributed to the plants’ capacity for absorption, transformation, and volatilization of these pollutants, thereby effectively eliminating or immobilizing them within the system. The process of plants removing metal elements in green roofs is shown in Figure 6.

Liu et al. conducted a comprehensive comparison between green roofs with vegetative cover and exposed substrates, revealing that the presence of plants reduced substrate loss by 5.14% (p < 0.05). Furthermore, it preserves the physical and chemical properties of the substrate, as well as the microbial conditions [44]. Maintaining adequate plant coverage and preventing substrate exposure is crucial for retaining nutrients in the substrate and enhancing microbial biomass in green roofs, thereby indirectly influencing water quality emissions from the substrate. Therefore, ensuring optimal plant coverage is imperative during green roof operation.

4.1.2. The Influence of a Singular Vegetation Type

The effluent water quality of systems is influenced by different plant types, with the effectiveness of pollutant removal in green roof runoff varying depending on the specific plant species [45]. Current research primarily focuses on comparing the respective contributions of various individual species in green roof ecosystems.

The study conducted by Thi-Dieu-Hien et al. involved a 30-day experiment on wetland roofs using eight different plant species [26]. Two distinct hydraulic loading rate (HLR) conditions were established, and Kyllinga brevifolia Rottb consistently exhibited the highest growth rates in terms of both fresh biomass and dry biomass. Kyllinga brevifolia Rottb, Cyperusjavanicus Houtt, and Imperata cylindrical ranked higher in terms of leaf area, indicating their potential to enhance urban green coverage. The results demonstrated a positive correlation between plant biomass and TN and TP removal rates, with the aforementioned three species exhibiting TN levels below 10 mg L⁻¹ in the effluent water, which is lower than other wetland roof systems. Regarding TP removal, all plants displayed similar phosphorus removal rates within the wetland roof system. Overall analysis suggests that Kyllinga brevifolia Rottb, Cyperusjavanicus Houtt, and Imperata cylindrical are considered to have significant greening effects and wastewater treatment capabilities.

Chai et al. conducted a three-year green roof experiment from July to September, focusing on the summer months and utilizing two distinct plant species [46]. The experimental findings revealed that the choice of plants exerted a significant influence on chemical oxygen demand (COD) levels (p < 0.01). Ophiopogon japonicus (L. f.) Ker Gawl. exhibited superior COD control compared to Ophiopogon japonicus ‘Nanus’. Notably, Ophiopogon japonicus (L. f.) Ker Gawl demonstrated remarkable environmental adaptability, while Ophiopogon japonicus ‘Nanus’ struggled with adaptation issues and displayed higher mortality rates. The absence of proper root fixation on green roofs resulted in reduced resistance against rainwater erosion, further compounded by soil subsidence and substrate blockage in lower layers, ultimately leading to a substantial increase in COD concentration.

Liu et al. conducted separate studies on the treatment efficiency of C4, C3, and CAM plants in green roofs for wastewater [47]. The research findings demonstrated that TN removal rates ranged from 65.26% to 90.52%, ${\mathrm{NO}}_3^{-}$-N removal rates ranged from 77.83% to 93.97%, ${\mathrm{NH}}_4^{+}$-N removal rates ranged from 83.32% to 96.31%, TP removal rates ranged from 93.77% to 98.94%, ${\mathrm{PO}}_4^{3-}$-P removal rates ranged from 96.36% to 99.43%, TSS removal rates ranged from 79.27% to 97.38%, and COD removal rates ranged from 79.94% to 98.92%. Moreover, a comparison revealed significantly higher TN, ${\mathrm{NO}}_3^{-}$-N, ${\mathrm{NH}}_4^{+}$-N, TP, ${\mathrm{PO}}_4^{3-}$-P, TSS, and COD removal efficiencies in C4 plants (Eremochloa ophiuroides and Cynodon dactylo) and C3 plants (Poa pratensis and Festuca arundinacea), compared with CAM plants (Sedum lineare and Callisia repens).

Gong et al. conducted a study on the purification effects of four plant species belonging to the Sedum genus (Sedum aizoon L., Sedum lineare Thunb., Sedum spurium cv. Coccineum, and Sedum spectabile) on green roofs [48]. The results revealed no significant variations in TN and ${\mathrm{NH}}_4^{+}$-N concentrations among different plant species. However, Liu et al. observed notable differences in the TN concentration between Sedum spectabile and Ophiopogon japonicas (Linn. f.) Ker-Gawl in their investigation, suggesting higher nitrogen utilization efficiency by plants within the Sedum genus [34]. Considering that all four plants utilized in Gong’s experiment belong to the same genus, it is more plausible that distinctions exist between plants from diverse families. Furthermore, there were no significant differences in TP concentrations among individual plant groups, providing support for the notion that vegetation types do not significantly impact TP absorption.

In a subsequent study, Thomaidi et al. investigated the effects of Atriplex halimus, Polygala myrtifolia, and Geranium zonale on greywater treatment in green roofs [49]. In a system filled with 10 cm of gravel without any plants, the average removal rate of COD was 70%. However, when Polygala myrtifolia and Atriplex halimus were introduced into the system as vegetation cover, the average COD removal rates significantly increased to 78% and 82%, respectively. Nevertheless, as the substrate thickness increased to 20 cm, the contribution of plants to organic matter removal became limited. Plant roots can enhance filtration by obstructing particulate matter and providing attachment surfaces for microorganisms to facilitate degradation [50]. However, the substrate in the vertical flow system also possesses the above-mentioned mechanisms, which can enhance treatment efficiency. It is evident that with increasing substrate thicknesses, the role of plants in removing organic matter and TSS diminishes. Therefore, for the effective removal of organic matter and TSS from water bodies on green roofs, plant presence is more crucial in extensive green roofs with smaller thicknesses compared to intensive green roofs with larger thicknesses. Additionally, in this study, TP removal was not influenced by plant presence since adsorption and chemical reactions between phosphorus and substrates are considered primary mechanisms for TP removal rather than plant uptake.

4.1.3. The Influence of Vegetation Combinations

In recent years, researchers have shifted their focus from studying the impact of individual plant species on the water quality of green roofs to investigating synergistic combinations of diverse plant species. The research findings demonstrate that mixed planting in green roof systems exhibits superior performance in terms of plant survival, canopy density, plant height, and growth when compared to single-species planting. Incorporating mixed planting on green roofs effectively enhances plants’ adaptability and resilience towards diverse climatic conditions [51].

Caceres et al. conducted a study to evaluate the green coverage and survival rate under mixed conditions on green roofs by selecting four species (Phyla nodiflora, Grindelia cabrerae, Eustachys retusa, and Sedum mexicanum) with different growth forms and stress resistance for mixed planting [52]. A total of 11 experimental groups were established using combinations of two, three, or four species. After one year, significant differences in both green coverage and survival rate were observed among different plant combinations, with nine combinations having a total green coverage of >80% and six combinations having a total survival rate of >80%. By the end of the second year, only five combinations (P. nodiflora and E. retusa; G. cabrerae and E. retusa; G. cabrerae, E. retusa and S. mexicanum; P. nodiflora, E. retusa and S. mexicanum; P. nodiflora, G. cabrerae, E. retusa and S. mexicanum) maintained a total green coverage and survival rate between 60 and 80%. The findings suggest that there are variations in green coverage and survival rates among different mixed planting combinations on green roofs. Therefore, when selecting plant combinations for green roofs, it is crucial to consider species persistence and colonization mechanisms as well as understand the spatial heterogeneity of plants to establish long-term stable plant diversity. Furthermore, increasing species richness and enhancing plant diversity can improve nitrogen retention in water quality purification systems, which is essential for maintaining water quality and sustaining ecosystem health [53].

Liao et al. established two types of green roofs: one with a predominant coverage of Phedimus kamtschaticus (Fisch.) t Hart, encompassing over 95% of the area, along with a small proportion of sedum sexangulare L. and sedum album L. (referred to as sedum mats), while the other group consisted of 14 species of non-grass herbaceous plants native to the eastern and central regions of North America (referred to as native plants) [54]. The study assessed the levels of TN, dissolved P, dissolved K, dissolved Ca, dissolved Na, and dissolved Mg loads. It was observed that compared to the group planted with native plants, the sedum mats group exhibited significantly lower nutrient load values (p < 0.05), resulting in reductions ranging from 21 to 64%. Furthermore, in comparison to the native plants group, the sedum mats group demonstrated decreased turbidity, EC, TSS loads, TDS concentrations, and load values, which were reduced by 41%, 19%, 32%, 19%, and 41%, respectively (p < 0.05). The pH emissions from the sedum mats group were higher than those emitted by the native plants group (7.39) (p < 0.05). These findings suggest that when it comes to reducing nutrient leaching from green roofs, the sedum mats group outperforms its counterpart planted with native plants. Possible reasons include, firstly, that sedum mats exhibit higher vegetation coverage, thereby enhancing nutrient absorption. They possess a greater capacity to intercept runoff, consequently reducing discharge levels to a certain extent. Secondly, the composition of vegetation within each group plays a significant role. Sedum mats comprise perennial succulents that thrive throughout the year; in contrast, native plants predominantly consist of annuals and short-lived perennials that perish during winter months. Plants with shorter lifespans tend to generate plant parts with elevated nitrogen and phosphorus concentration [55], which can result in increased nutrient concentration upon decomposition. Thirdly, the dense root network of sedum mat plants effectively impedes particulate matter and minimizes particle loss on green roofs.

Hu et al. established four types of mixed turf grasses, consisting of the following specific types and ratios: Group 1—Cynodon dactylon: Zoysia japonica: Lolium perenne = 7:2:1; Group 2—Poa pratensis: Agrostis matsumurae: Lolium perenne = 5:4:1; Group 3—Poa pratensis: Festuca elata: Lolium perenne = 5:3:2; Group 4—Zoysia japonica: Cynodon dactylon = 2:1 [56]. Simultaneously, the experiment employed four distinct substrate proportions denoted as groups A, B, C, and D. By utilizing an orthogonal design approach, a total of 16 experimental groups were established through the combination of these four plant combinations and substrates. The ASFV method was employed to comprehensively assess the removal efficiency of ${\mathrm{NH}}_4^{+}$-N, SS, COD, TP, and TN for each combination and prioritize them accordingly. The results revealed significant variations in purification effects among different plant combinations at the same substrate level. Combination A2 exhibited a superior stormwater runoff purification effect, while combination A3 ranked fifteenth out of all sixteen tested combinations. In substrate A and D groups, plants in Group 2 showed the best purification effect. In the substrate C group, only when combined with plants in Group 3, can a better water purification effect be achieved. Therefore, the water purification process is significantly influenced by different combinations of plants. However, when considering various plant combinations, it is imperative to also account for the interaction between mixed planting and diverse substrate types in order to attain optimal effects on water purification.

Table 1. Plant species used for green roof planting in recent years.

Plants	References
Ophiopogon japonicus (L. f.) Ker Gawl., Ophiopogon japonicus ‘Nanus’	[46]
Cynodon dactylon, Cyperus javanicus Houtt Cyperus rotundus L., Eleusine indica (L.) Gaertn Imperata cylindrical, Kyllinga brevifolia Rottb Struchium sparganophorum (L.) Kuntze, Zenith zoysia grass	[26]
Hylotelephium erythrostictum (Miq.) H. Ohba, Iris tectorum Maxim. Ophiopogon japonicus (L. f.) Ker Gawl.	[34,45]
Callisia repens L., Cynodon dactylon (L.) Persoon Eremochloa ophiuroides (Munro) Hack. Festuca arundinacea Schreb., Poa pratensis L., Sedum lineare Thunb.	[47]
Sedum aizoon L., Sedum lineare Thunb. Sedum spurium cv.Coccineum, Sedum spectabile	[48]
Briza maxima, Conyza sp.Digitaria sanguinalis, Dittrichia viscosa Filago pyramidata, Gomphocarpus fruticosus Illecebrum verticillatum, Lavandula stoechas subsp. luisieri Pleurochaete squarrosa, Sedum sediforme Teucrium scorodonia, Trifolium angustifolium, Vulpia geniculata	[57]
Axonopus Compressus, Wedelia Trilobata	[58]
Atriplex halimus, Geranium zonale, Polygala myrtifolia	[49]
Agrostis matsumurae, Cynodondactylon, Festuca elata Lolium perenne, Poa pratensis, Zoysia japonica	[56]
Eustachys retusa, Grindelia cabrerae Phyla nodiflora, Sedum mexicanum	[52]

Table 1. Plant species used for green roof planting in recent years.

Plants	References
Ophiopogon japonicus (L. f.) Ker Gawl., Ophiopogon japonicus ‘Nanus’	[46]
Cynodon dactylon, Cyperus javanicus Houtt Cyperus rotundus L., Eleusine indica (L.) Gaertn Imperata cylindrical, Kyllinga brevifolia Rottb Struchium sparganophorum (L.) Kuntze, Zenith zoysia grass	[26]
Hylotelephium erythrostictum (Miq.) H. Ohba, Iris tectorum Maxim. Ophiopogon japonicus (L. f.) Ker Gawl.	[34,45]
Callisia repens L., Cynodon dactylon (L.) Persoon Eremochloa ophiuroides (Munro) Hack. Festuca arundinacea Schreb., Poa pratensis L., Sedum lineare Thunb.	[47]
Sedum aizoon L., Sedum lineare Thunb. Sedum spurium cv.Coccineum, Sedum spectabile	[48]
Briza maxima, Conyza sp.Digitaria sanguinalis, Dittrichia viscosa Filago pyramidata, Gomphocarpus fruticosus Illecebrum verticillatum, Lavandula stoechas subsp. luisieri Pleurochaete squarrosa, Sedum sediforme Teucrium scorodonia, Trifolium angustifolium, Vulpia geniculata	[57]
Axonopus Compressus, Wedelia Trilobata	[58]
Atriplex halimus, Geranium zonale, Polygala myrtifolia	[49]
Agrostis matsumurae, Cynodondactylon, Festuca elata Lolium perenne, Poa pratensis, Zoysia japonica	[56]
Eustachys retusa, Grindelia cabrerae Phyla nodiflora, Sedum mexicanum	[52]

4.2. Substrate

As a crucial component of green roofs, substrates have a significant impact on the water quality of green roof runoff. It is widely recognized that substrates possess adsorption and filtration capabilities, enabling the direct absorption and filtration of nitrogen, phosphorus, chemical oxygen demand, suspended solids, etc., thereby reducing water pollutants. Additionally, microorganisms attached to the substrate contribute to biodegradation and pollutant removal from the water. However, certain studies have indicated that green roof substrates may also act as sources of pollutants due to nutrient leaching and eutrophication processes, leading to adverse effects on runoff water quality. Moreover, substrates indirectly influence plant growth and function while affecting water quality as well. Depending on the type of growing substrate used [59], green roofs can serve both as sources and sinks for pollutants. Therefore, when constructing green roofs, careful consideration should be given to factors such as the choice of materials, appropriate proportions, substrate thickness, and the presence of amendment. The composition of green roof substrate components and their substrate depth settings in rainwater and sewage treatment applications from 2014 to 2024 are presented in Table 2.

4.2.1. Substrate Composition and Proportion

The study suggests that the substrate components can be categorized into inorganic and organic constituents [60]. The inorganic components serve as the source of mineral elements, which are beneficial for enhancing cation exchange and providing trace nutrients to plants. The organic components also provide nutrition for plant growth [61]. In the design of actual substrates, it is common to have the simultaneous presence of both inorganic and organic components. Given the variations in physical properties and chemical characteristics among different components, it is typically imperative to experimentally determine the proportions and combinations for achieving optimal performance in green roof treatments.

In recent years, Peczkowski et al. have developed two types of substrates for green roofs: one based on Lightweight Expanded Clay Aggregate (LECA) consisting of 60% horticultural soil, 20% sand, and 20% LECA (4–8 mm in size), and the other based on perlite comprising 60% horticultural soil, 20% sand, 15% perlite, and 5% LECA (4–8 mm in size) [62]. The study investigated various parameters, including TN, ${\mathrm{NO}}_3^{-}$-N, ${\mathrm{NO}}_2^{-}$-N, ${\mathrm{NH}}_4^{+}$-N, TP, ${\mathrm{PO}}_4^{3-}$-P, as well as heavy metals such as copper (Cu), zinc (Zn), lead (Pb), and cadmium (Cd). The findings suggest that no substantial enhancement was observed in the water quality of green roofs. The concentrations of TN, copper (Cu), and zinc (Zn) in the LECA substrate and perlite substrate were found to be significantly higher than those in rainwater, with a notable increase observed specifically in the concentration of the perlite substrate.

The composition of green roof substrates (13 samples) and commonly used mineral compounds (29 samples of building aggregates) in terms of phosphorus (P), copper (Cu), nickel (Ni), cadmium (Cd), and zinc (Zn) concentration was investigated by Karczmarczyk et al. [63]. The concentration of P, Cu, Ni, Cd, and Zn in the runoff water from green roofs was also determined. The results revealed a positive correlation between the metal concentration in substrate composition and the quality of runoff water. In the study, with the exception of one group where substrate samples had previously received fertilization, disregarding the fertilizer factor, the findings suggest that potential pollution in green roofs is associated with the specific substrate chosen and its composition of compounds. Natural materials such as sand and gravel, as well as artificial materials like expanded clay and ash and crushed red brick, are considered to potentially contribute to higher levels of phosphorus pollution. Sand and crushed red brick may result in increased leaching of nickel concentration.

Rey et al. employed a combination of organic components, including compost (C), coco-peat (CP), rice husk (R), and humic soil (So), along with inorganic components such as expanded clay (ECl), perlite (P), coarse pumice (Pu), sand (S), and zeolite (Z) materials, to design various volume ratios of substrates [64]. These substrates were compared against commercially available extensive and intensive substrates. The findings revealed that the substrate (So20:ECl10:Pu40:S10:P10:Z10) and the substrate (CP20:ECl5:Pu60:S5:P5:Z5) exhibited favorable physical characteristics, including low bulk density and high water retention capacity, which supported normal plant growth. Paepalanthus alpinus plants demonstrated a 100% survival rate when grown in these two substrates. In comparison to the commercial extensive substrate, both mixed substrates showed significantly reduced concentrations of TKN, ${\mathrm{PO}}_4^{3-}$-P, TSS, turbidity, COD, BOD, and coliforms; however, they still acted as sources for these pollutants when compared to rainwater inflow. No significant difference was observed in TP, ${\mathrm{NO}}_3^{-}$-N, and ${\mathrm{NO}}_2^{-}$-N between the effluent concentrations of the mixed substrates and influent concentrations. The modified mixed substrates displayed lower pollutant concentrations compared to the commercial extensive substrate, suggesting their effectiveness in reducing runoff pollution.

The leaching of nutrients and organic matter from the substrate can result in the eutrophication of water and an elevation in pollutant concentrations in the effluent, potentially surpassing those found in rainwater. Therefore, to mitigate the substrate’s potential as a source of pollutants, careful selection of appropriate substrate materials and proportions is crucial to enhance both pollutant degradation and retention capacity within the substrate.

The substrate design by Vijayaraghavan and Raja involved varying proportions of vermiculite, perlite, crushed brick, sand, and coco-peat [65]. Out of the 18 designs tested, the mixture consisting of 20% vermiculite, 30% perlite, 20% crushed brick, 10% sand, and 20% coco-peat exhibited superior characteristics compared to other combinations and individual media groups. This particular composition demonstrated a bulk density of 431 kg m⁻³, an air-filled porosity of 19.5%, a hydraulic conductivity reaching up to 4570 mm h⁻¹, and a water holding capacity of up to 39.4%. Moreover, Portulaca grandiflora plants cultivated in this mixed substrate displayed optimal growth performance with a biomass increase of approximately 380%. Furthermore, this specific mixture composition showcased an effective removal rate exceeding 97% for heavy metal ions (Al, Cd, Cr, Cu, Fe, Ni, Pb, and Zn).

Afterward, Vijayaraghavan and Badavane employed varying volume ratios of Purosil (a processed siliceous soil), vermiculite, Sand, LECA (Lightweight Expanded Clay Aggregate), coco-peat, and Sargassum wightii (commonly known as Sargassum seaweed) for the substrate design [66]. The findings revealed that out of the 13 different ratio designs tested, the optimal combination for the substrate mixture consisted of 20% Purosil, 30% vermiculite, 10% sand, 20% LECA, 10% coco-peat, and 10% S. wightii. This blend exhibited a bulk density of 495 kg m^−³ with an air-filled porosity of 21%, a hydraulic conductivity reaching up to 5524 mm hour⁻¹, and a water holding capacity of up to 67.6%. Moreover, the Portulaca grandiflora plants cultivated on this substrate demonstrated robust growth performance, with an approximately 2.72-fold increase in biomass over a period of 40 days of operation. Furthermore, the mixed substrates composed of these six materials at different volume ratios also displayed remarkable binding capacities towards heavy metals (Al, Cd, Cr, Cu, Fe, Ni, Pb, and Zn), with removal rates surpassing 93.7%.

4.2.2. Substrate Thickness

When constructing green roof substrates, it is crucial to consider not only the composition and proportion of the substrate but also its optimal thickness. The thickness of the substrate plays a significant role in water quality improvement by influencing plant and microbial growth as well as functionality. A thicker substrate prolongs water passage, thereby increasing contact time between pollutants and the substrate, plant roots, and microorganisms. This extended contact enhances pollutant removal efficiency. Moreover, a thicker substrate with more pores and a larger surface area facilitates improved filtration and sedimentation of particulate matter. Additionally, it supports higher richness and diversity of microbial attachment growth, which efficiently transforms nutrients and decomposes organic matter for effective pollutant treatment. Furthermore, plants’ roots can fully develop in a thicker substrate, enabling them to absorb more nutrients directly while providing robust physical filtering at root zones along with larger surfaces for microbial growth. Considering these comprehensive factors mentioned above, it is evident that a thicker substrate often yields superior treatment results; however, system operation costs and overall design should also be taken into account.

Chai et al. utilized perlite and recycled bricks to establish green roofs with substrate thicknesses of 10 cm and 20 cm, respectively [46]. The findings indicated that the adsorption capacity of suspended solids (SS) was influenced by the substrate thickness, with a thicker substrate exhibiting greater SS adsorption ability and reducing the concentration of SS in the effluent. Furthermore, substrate thickness had a significant impact on ${\mathrm{NH}}_4^{+}$-N concentration in the effluent (p < 0.05). Increasing the substrate depth from 10 cm to 20 cm resulted in a decreased ${\mathrm{NH}}_4^{+}$-N concentration, indicating that increasing the substrate depth can better regulate ${\mathrm{NH}}_4^{+}$-N through adsorption, retention, and transformation processes. However, this study did not observe any significant effect of substrate thickness on TN, TP, and COD concentrations.

In their study, Gong et al. investigated the efficacy of green roofs with varying substrate thicknesses (10 cm, 15 cm, and 20 cm) and a module area of 0.5 square meters in treating runoff water quality [16]. The findings revealed a gradual reduction in average concentrations of TN, ${\mathrm{NO}}_3^{-}$-N, and ${\mathrm{NH}}_4^{+}$-N in the effluent as the substrate thickness increased.

The study conducted by Thomaidi et al. employed perlite and vermiculite as substrates for the establishment of green roofs, with substrate thicknesses of 10 cm and 20 cm, respectively [49]. The findings revealed that when the substrate thickness was increased to 20 cm, both types of substrates demonstrated enhanced removal efficiencies for TSS, Turbidity, BOD, and COD, achieving optimal removal rates of 93%, 93%, 91%, and 91%, respectively. However, a notable decline in removal efficiency was observed when the substrate thickness decreased to 10 cm, resulting in removal rates ranging from 60% to 75% for the aforementioned indicators.

The study conducted by Park et al. investigated the impact of varying substrate thicknesses on effluent quality in green roofs, employing substrate thicknesses of 10 cm, 20 cm, and 40 cm [24]. The findings revealed a negative correlation between substrate thickness and heavy metal concentrations (Cu, Zn, Mn, and Cd) in roof runoff, particularly for Cu and Zn. Notably, an increase in substrate thickness was observed to significantly reduce the concentrations of these metals.

4.2.3. Substrate Amendment

Incorporating soil amendments into the substrate can enhance the development of aggregate structures (AS), which serve as fundamental components of soil. Well-developed AS possess excellent water retention capacity and mitigate nutrient leaching. Amendments are recognized as effective measures for controlling nutrient leaching, and their application in green roof substrates can contribute to controlling pollution.

When utilized as a soil amendment, biochar has the potential to enhance the absorption of both inorganic and organic pollutants while mitigating the leaching of nitrogen and phosphorus from the soil. In recent years, researchers have employed biochar on green roofs to investigate its impact on runoff quantity and quality. However, studies have yielded divergent conclusions regarding the effects of biochar on nutrient concentrations in green roof runoff. Some studies suggest that biochar can reduce nutrient concentrations in runoff [59,67], while others indicate that it may release nutrients during operation, resulting in elevated nutrient levels in the runoff. This phenomenon is attributed to the fact that biochar is derived from mineral-rich sewage sludge [68]. Researchers have conducted a comprehensive investigation on green roofs, examining the impact of three different sources of biochar, namely wood biochar, sewage sludge biochar, and food waste biochar [69]. The incorporation of wood biochar significantly influences the effluent quality of green roofs by reducing average concentrations of ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_3^{-}$-N, TP, COD, and BOD₅ by 63%, 4%, 13.4%, 32.7%, and 4.7%, respectively, whereas the average concentrations of metals As, Ca, Cd, Cr, Mg, Ni, and Zn decrease within a range of 6.9–99%. However, TN concentration experiences an average increase of 527%. Average concentrations of metals Cu, Hg, K, and Pb exhibit an increase within a range between 8.3 and 325.5%. These findings suggest that while wood biochar demonstrates positive effects in mitigating specific pollutants, it also acts as a source of certain contaminants. In a separate investigation conducted by Xiong et al., maize straw biochar (MSB) and rice husk biochar (RHB) were employed as materials [70]. The findings demonstrated that in comparison to rice husk biochar, corn straw biochar exhibited superior efficacy in reducing TN, ${\mathrm{NO}}_3^{-}$-N, and DOC concentrations. Both types of biochar significantly elevated the levels of TP and ${\mathrm{PO}}_4^{3-}$-P in runoff water; however, the impact on phosphorus elements was relatively less pronounced for rice husk biochar. Furthermore, this study investigated the effects of varying ratios of added biochar on runoff water with three ratios set at 10%, 15%, and 20%, respectively. The results revealed a significant decrease in TN and ${\mathrm{NO}}_3^{-}$-N concentrations in runoff water as the ratio of added biochar increased while TP and ${\mathrm{PO}}_4^{3-}$-P concentrations gradually rose.

The granulation process and particle size of biochar have significant impacts on stormwater runoff from green roofs. Granulated biochar can enhance plant growth, leading to increased leaf area and final biomass [71]. Moreover, the use of granulated biochar reduces TSS concentration and improves the water quality index (WQI). Smaller particle-sized biochar is more effective in reducing particle loss and nutrient leaching compared to larger particles due to its ability to form water-stable aggregates and stronger water retention capacity [42].

In addition to utilizing biochar as an amendment, researchers are also investigating the utilization of alternative substances to enhance green roof performance. Zhang et al. conducted a study on the pollution control capability of polyaluminium chloride (PAC) and bentonite when introduced into runoff water from green roofs [72]. The findings revealed that both amendments compromised the ability of green roofs to regulate ${\mathrm{NH}}_4^{+}$-N. Various concentrations of PAC and bentonite were examined, demonstrating that green roofs with PAC consistently exhibited superior pollutant control compared to those with bentonite. Specifically, the green roof incorporating 2.0% PAC displayed an increase in removal rates for ${\mathrm{NO}}_3^{-}$-N, TN, and TP by 204.50%, 148.36%, and 38.00%, respectively, in comparison to the group without any amendment. Expanding upon this research, Fei et al. further investigated the treatment performance of green roofs supplemented with additions of 2% polyaluminium chloride (PAC), polyferric sulfate (PFS), polyvinyl alcohol (PVA), methylcellulose (MC), carboxymethyl cellulose sodium (CMC), and hydroxypropyl methylcellulose (HPMC) [73]. The results indicated that PVA readily formed unstable aggregates, leading to the leaching of N and P nutrients. Additionally, the inclusion of MC, CMC, or HPMC did not significantly enhance pollutant interception capabilities either; thus, it is not recommended to employ these additives at a concentration level of 2%. Both PAC and PFS facilitated bridging adsorption and coagulation between particles while improving retention capacity for pollutants within the substrate layer. However, it should be noted that incorporating PAC may result in aluminum contamination, which inhibits plant growth and pollutes the environment; therefore, PFS appears to be a more suitable choice.

Table 2. Different green roof main substrate component and depth settings.

Main Substrate Component	Depth	References
Expanded clay, Spongilite, Peat, Brick rubble, Biochar from wood, Biochar from sewage sludge, Biochar from food waste, and Dried sewage sludge	150 mm	[69]
Expanded clay, Crushed marl, Peat, Recycled bricks, Biochar	100 mm	[74]
Peat soil, vermiculite	250 mm	[73]
Fractured tiles, Red lava, Fine pumice, Compost, Peat, Sand, Coconut fiber, Gravel	60 mm, 90 mm	[75]
Soil, Rice husk biochar, Maize stalk biochar, Perlite, Vermiculite	100 mm	[70]
Soil, Cocopeat, Loofah, Perlite	80 mm	[76]
Compost, Paper sludge, Pelletized paper sludge, Vulcaflor	100 mm	[77]
Unprocessed biochar, Granulated biochar	80 mm	[42]
Perlite, Vermiculite, LECA	150 mm, 250 mm	[49]
Biochar, Vermiculite, Porous aggregates, Composted organic matter, Fine sand	80 mm	[71]
Perlite, Peatmoss, Vermiculite	100 mm, 200 mm, 400 mm	[24]
Rural soil, Peat soil, Pine needle, Perlite, Vermiculite	50 mm, 100 mm, 150 mm	[45]
Peat soil, Vermiculite, Perlite, Biochar, Sawdust	50 mm, 100 mm, 150 mm, 200 mm	[78]
Wheat straw	>200 mm	[79]
Sand, Gravel, Limestone, Lightweight Aggregates, Expanded clay and ash, Crushed red brick	-	[63]
Sand, Gravel, Brick, Rubble, Bark, Peat, Compost, Polonite	120 mm	[80]
Humic soil, Compost, Coco-peat, Rice husk, Coarse pumice, Expanded clay, Sand, Zeolite, Perlite	100 mm	[64]
Horticultural soil, Sand, Expanded clay aggregate, Light expanded clay aggregate, Perlite	80 mm	[62]
Waste building material substrate, Local natural soil	200 mm, 250 mm, 300 mm	[81]
Stabilized sludge, Biochar, Pumice, Wood chips, Topsoil, Controlled release fertilizer	100 mm, 150 mm	[82]
Rural soil, Peat soil, Pine needle, Perlite, Vermiculite	50 mm, 100 mm, 150 mm	[34]
Pastoral soil, Turfy soil, Pine needles	50 mm, 100 mm, 150 mm	[16]
Peat, Vermiculite, Perlite, Sawdust, Biochar	100 mm	[59]
Rural soil, Peat soil, Pine needles, Perlite, Vermiculite	50 mm, 100 mm, 150 mm	[83]
Modified perlite, Modified recycled bricks,	100 mm, 200 mm	[46]
Perlite, Coal ash	200 mm	[84]
Local soil, Peat soil, Vermiculite, Perlite	50 mm, 100 mm	[85]
Pumice, Lava, Perlite, Activated charcoal, Zeolite	50 mm, 100 mm, 150 mm	[21]
Peat, Volcanic rock, Wood biochar, Olive husk biochar	200 mm	[86]
Expanded clay, Granulated cork, Organic matter from urban solid waste compost, Crushed egg shell	150 mm	[87]
Crushed bark, Sphagnum moss, Compost, Recycled, Crushed brick, Biochar from Birch Wood	30 mm, 40 mm	[68]
Crushed, Recycled brick, Compost, Crushed bark	50 mm	[88]
Purosil, Vermiculite, Sand, Lightweight expanded clay aggregates, Coco-peat, Sargassum wightii	100 mm	[66]
Expanded slate, Compost	10 mm	[89]
Peat soil, Vermiculite, Perlite, Sawdust,	150 mm	[90]
Vermiculite, Perlite, Crushed brick, Sand, Coco-peat	100 mm	[65]

Table 2. Different green roof main substrate component and depth settings.

Main Substrate Component	Depth	References
Expanded clay, Spongilite, Peat, Brick rubble, Biochar from wood, Biochar from sewage sludge, Biochar from food waste, and Dried sewage sludge	150 mm	[69]
Expanded clay, Crushed marl, Peat, Recycled bricks, Biochar	100 mm	[74]
Peat soil, vermiculite	250 mm	[73]
Fractured tiles, Red lava, Fine pumice, Compost, Peat, Sand, Coconut fiber, Gravel	60 mm, 90 mm	[75]
Soil, Rice husk biochar, Maize stalk biochar, Perlite, Vermiculite	100 mm	[70]
Soil, Cocopeat, Loofah, Perlite	80 mm	[76]
Compost, Paper sludge, Pelletized paper sludge, Vulcaflor	100 mm	[77]
Unprocessed biochar, Granulated biochar	80 mm	[42]
Perlite, Vermiculite, LECA	150 mm, 250 mm	[49]
Biochar, Vermiculite, Porous aggregates, Composted organic matter, Fine sand	80 mm	[71]
Perlite, Peatmoss, Vermiculite	100 mm, 200 mm, 400 mm	[24]
Rural soil, Peat soil, Pine needle, Perlite, Vermiculite	50 mm, 100 mm, 150 mm	[45]
Peat soil, Vermiculite, Perlite, Biochar, Sawdust	50 mm, 100 mm, 150 mm, 200 mm	[78]
Wheat straw	>200 mm	[79]
Sand, Gravel, Limestone, Lightweight Aggregates, Expanded clay and ash, Crushed red brick	-	[63]
Sand, Gravel, Brick, Rubble, Bark, Peat, Compost, Polonite	120 mm	[80]
Humic soil, Compost, Coco-peat, Rice husk, Coarse pumice, Expanded clay, Sand, Zeolite, Perlite	100 mm	[64]
Horticultural soil, Sand, Expanded clay aggregate, Light expanded clay aggregate, Perlite	80 mm	[62]
Waste building material substrate, Local natural soil	200 mm, 250 mm, 300 mm	[81]
Stabilized sludge, Biochar, Pumice, Wood chips, Topsoil, Controlled release fertilizer	100 mm, 150 mm	[82]
Rural soil, Peat soil, Pine needle, Perlite, Vermiculite	50 mm, 100 mm, 150 mm	[34]
Pastoral soil, Turfy soil, Pine needles	50 mm, 100 mm, 150 mm	[16]
Peat, Vermiculite, Perlite, Sawdust, Biochar	100 mm	[59]
Rural soil, Peat soil, Pine needles, Perlite, Vermiculite	50 mm, 100 mm, 150 mm	[83]
Modified perlite, Modified recycled bricks,	100 mm, 200 mm	[46]
Perlite, Coal ash	200 mm	[84]
Local soil, Peat soil, Vermiculite, Perlite	50 mm, 100 mm	[85]
Pumice, Lava, Perlite, Activated charcoal, Zeolite	50 mm, 100 mm, 150 mm	[21]
Peat, Volcanic rock, Wood biochar, Olive husk biochar	200 mm	[86]
Expanded clay, Granulated cork, Organic matter from urban solid waste compost, Crushed egg shell	150 mm	[87]
Crushed bark, Sphagnum moss, Compost, Recycled, Crushed brick, Biochar from Birch Wood	30 mm, 40 mm	[68]
Crushed, Recycled brick, Compost, Crushed bark	50 mm	[88]
Purosil, Vermiculite, Sand, Lightweight expanded clay aggregates, Coco-peat, Sargassum wightii	100 mm	[66]
Expanded slate, Compost	10 mm	[89]
Peat soil, Vermiculite, Perlite, Sawdust,	150 mm	[90]
Vermiculite, Perlite, Crushed brick, Sand, Coco-peat	100 mm	[65]

4.3. Slope

Unlike artificial wetlands or vertical green walls, green roofs are often not positioned vertically due to their unique location and may have varying slopes. Therefore, when designing green roofs, it is crucial to consider the variable roof slope in order to accurately evaluate the performance of green roof treatments [91].

Beecham and Razzaghmanesh conducted an investigation on the water purification effects of green roofs with slopes of 1° and 25° [92]. The results revealed no significant variations in pH, turbidity, ${\mathrm{NO}}_3^{-}$-N, ${\mathrm{NO}}_2^{-}$-N, ${\mathrm{NH}}_4^{+}$-N, potassium (K), sodium (Na), calcium (Ca), and magnesium (Mg) concentrations between the two slopes when considering the presence of vegetation cover as well as identical substrate type and thickness (p > 0.05). These findings suggest that factors related to vegetation and substrate may exert a more substantial influence on water quality compared to slope factors.

The study conducted by Castro et al. examined the impact of green roofs with slopes of 0° and 15° on runoff water quality [93]. The research findings revealed no significant differences in TN, ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_3^{-}$-N, TP, pH, and turbidity concentration between the two slope treatments with vegetation cover. However, green roofs with a 15° slope exhibited slightly higher TSS, BOD₅, and COD concentrations compared to the group treated with a 0° slope. This observation can be attributed to the steeper incline of the 15° slope, which enhances rainwater flow and erosion capacity, facilitating the transport of solid particles and organic matter from the substrate.

In their study, Liu et al. conducted experiments on green roofs with varying slopes of 2%, 7%, and 12% [45]. The findings revealed a positive correlation between the slope of the green roof and the concentrations of F- and TP in water. Specifically, the F- concentration was significantly lower on the green roof with a slope of 2% compared to those with slopes of 7% and 12% (p < 0.01), while the TP concentration was significantly lower on the green roof with a slope of 7% compared to that on the green roof with a slope of 12% (p < 0.05). However, no significant differences were observed among the three slopes regarding TN, ${\mathrm{NH}}_4^{+}$-N, ${\mathrm{NO}}_2^{-}$-N, Cl⁻, ${\mathrm{SO}}_4^{2-}$, pH, EC, ESP, and TSS concentration.

4.4. Operating Conditions

Different operating conditions, such as hydraulic load rate (HLR) and hydraulic retention time (HRT), exert specific influences on the treatment performance of green roofs. Excessively high HLR in the system can detrimentally impact filtration rates, while insufficient HRT diminishes water-plant and substrate contact time. Hence, when designing green roofs for water treatment purposes, it is imperative to establish reasonable and effective operating conditions.

4.4.1. Hydraulic Retention Time

The study conducted by Xu et al. investigated the treatment efficiency of a hydroponic green roof system under three different hydraulic retention times, namely 4 days, 6 days, and 8 days [27]. The findings revealed that at a hydraulic retention time of 4 days, the system exhibited an average turbidity removal rate of 67.4%, which further increased to 80.0% when the retention time was extended to 8 days. However, it is noteworthy that at a retention time of 6 days, the effluent turbidity surpassed the influent turbidity levels. In terms of chemical oxygen demand (COD) removal rates, values were determined as follows: for hydraulic retention times of 4, 6, and 8 days, respectively –69.5%, 66.0%, and 81%. Similarly, BOD₅ removal rates were observed to be 69.4%, 57.1%, and 97%. Regarding anionic surfactant (MBAS concentration indication), its elimination rates were recorded as 22.8%, 31.4%, and 88%. The extension of the appropriate hydraulic retention time leads to a significant reduction in turbidity, organic matter, and anionic surfactant concentration, thereby enhancing the quality of wastewater.

4.4.2. Hydraulic Load Rate

As the hydraulic load ratio increases, it will have a certain impact on the growth condition and survival rate of plants, as well as limit their functionality. Thi-Dieu-Hien et al. investigated the influence of wetland roofs on septic tank wastewater purification under two different hydraulic load rates (HLR1: 288 ± 19 m³ ha⁻¹ day⁻¹; HLR2: 394 ± 13 m³ ha⁻¹ day⁻¹) [26]. Under HLR1, all plants exhibited normal survival rates. However, with an increase in hydraulic load ratio from HLR1 to HLR2, plant growth was delayed, and some plants displayed yellowing leaves or even mortality. The COD removal efficiency at HLR1 ranged from 16 to 30%, with a removal rate of 67–86 kg ha⁻¹ day⁻¹, whereas at HLR2, the COD removal efficiency ranged from 27 to 33%, with a removal rate of 61–79 kg ha⁻¹ day⁻¹. The results revealed that although HLR1 exhibited a lower COD removal rate compared to HLR2, it demonstrated a higher removal rate than HLR2. This discrepancy can be attributed to the limited absorption and decomposition of organic matter by plants and microorganisms due to the relatively high hydraulic load rates, which result in shorter retention times. Bui Xuan et al. also conducted experiments using four different hydraulic load rates (HLR1: 137 ± 6 m³ ha⁻¹ day⁻¹; HLR2: 210 ± 7 m³ ha⁻¹ day⁻¹; HLR3: 338 ± 9 m³ ha⁻¹ day⁻¹; HLR4: 456 ± 6 m³ ha⁻¹ day⁻¹) to investigate the treatment performance of wetland roofs for septic tank effluent [94]. When the Hydraulic Load Ratio is HLR1, HLR2, HLR3, and HLR4, the ratios of nitrogen uptake by plants to nitrogen removal by the system are 0.09, 0.10, 0.13, and 0.11, respectively. The ratios of phosphorus uptake by plants to phosphorus removal by the system are 0.47, 0.50, 0.68, and 0.53, respectively. This indicates that when the hydraulic loading rate is too high, plants’ ability to absorb nitrogen and phosphorus is also inhibited.

4.4.3. Water Feeding Patterns

The study conducted by Nguyen et al. investigated the impact of two water feeding patterns, namely continuous and intermittent, on the performance of wetland roof wastewater treatment [58]. Under identical substrate conditions and plant settings, the intermittent inflow method exhibited significantly higher efficiency in removing COD (62–64%) compared to the continuous inflow method (52–54%). Similarly, in terms of TN removal efficiency, the intermittent inflow method (80–87%) outperformed the continuous inflow method (73–80%). This can be attributed to enhanced oxygen diffusion into the system facilitated by the intermittent inflow method, which promotes nitrification and subsequently improves ammonium nitrogen removal. Consequently, it is evident that implementing an intermittent inflow approach can considerably enhance both COD and TN removal efficiencies in wetland roof systems; however, no significant impact was observed on TP removal.

4.4.4. Other Additional Conditions

The greywater treated by green roofs does not meet the standards for indoor non-potable use, and chlorination is considered a crucial step in the reuse of greywater. Petousi et al. integrated green roofs with chlorination technology to eliminate pathogens from greywater [95]. Their study demonstrates that when the storage period is less than 24 h, a chlorine dosage ranging from 3 to 7 mg L⁻¹ can be added to the effluent to ensure water quality within microbial standards. However, if the storage period exceeds 24 h, there is a significant regeneration of pathogenic microorganisms. In such cases, adding a chlorine dosage of 7 mg L⁻¹ can guarantee the complete inactivation of pathogenic microorganisms within three days for indoor non-potable use.

To address the issue of elevated total coliform concentrations, it is crucial to enhance filtration and disinfection treatment following green roof runoff management [96]. This strategic approach can effectively enhance the safety and accessibility of green roof runoff, thereby facilitating optimal utilization of rainwater resources.

4.5. Time

The newly established or installed green roofs may become a source of water pollutants, as the organic matter in these new green roofs is prone to decomposition or requires initial fertilization. Whether green roofs act as sinks or sources of pollutants depends not only on the substrate but also on their age [59].

Due to plant growth and substrate erosion, the water quality of green roofs may undergo temporal changes. Gong et al. conducted a two-year monitoring study on the water quality of a green roof from 2012 to 2013 [97]. The results demonstrated that after one year of operation, there was a significant decrease in nitrogen and phosphorus concentration in the outflow, both being lower than rainwater concentrations. Furthermore, it was observed that the green roof exhibited acid rain neutralization capabilities, pH stabilization, and reduction in turbidity and COD concentrations, leading to substantial enhancement of water quality.

The study conducted by Speak et al. compared the runoff water quality between a 43-year-old green roof and an adjacent traditional roof surface [98]. The findings demonstrated that the green roof functioned as a sink for ${\mathrm{PO}}_4^{3-}$-P and ${\mathrm{NO}}_3^{-}$-N, effectively removing these pollutants from the runoff. However, monitoring revealed higher concentrations of plumbum (Pb) in the green roof runoff compared to rainwater, indicating that green roofs may contribute to Pb pollution in water bodies. Therefore, it is crucial to carefully consider plant and substrate selection while establishing green roofs for purification and restoration purposes, particularly considering the potential for aging green roofs to become sources of metal contaminants.

Harper et al. conducted a nine-month study investigating the impact of green roofs on runoff water quality [91]. As the duration of operation increased, there was a gradual decrease in the concentrations of TN and TP in the effluent from the green roof. The initial concentration of TN, which exceeded 60 mg L⁻¹, decreased to approximately 10 mg L⁻¹, while TP concentration reduced from an initial value exceeding 30 mg L⁻¹ to approximately 5 mg L⁻¹. Furthermore, TOC levels in the system started at an initial concentration of 500 mg L⁻¹ and gradually declined to 50 mg L⁻¹ after several weeks, indicating a progressive reduction in organic matter dissolution and loss within the green roof with increasing operational time.

Akther et al. conducted laboratory and field investigations to examine nutrient leaching from green roofs [99]. The findings revealed an initial higher level of nutrient leaching, particularly nitrogen and phosphorus, during the early stages of green roof operation. However, with a prolonged duration of operation, there was a general decrease in the extent of nutrient leaching. This decline can be attributed to a reduction in available exchangeable nutrients within the substrate over time, resulting in diminished nutrient leaching. Consequently, long-term operation may contribute to an enhancement in water quality originating from green roofs.

4.6. Weather

Meteorological conditions, such as temperature, humidity, and precipitation, can all have an impact on the water quality of green roofs. Temperature and humidity play a crucial role in plant growth and physiological processes, as well as microbial activity, which subsequently affects water quality. The frequency and intensity of precipitation also influence both the runoff volume and water quality of green roofs; higher frequencies and intensities may result in increased pollutant transport from the substrate. Due to seasonal variations in temperature, humidity, and precipitation patterns, there is often a discernible seasonal fluctuation in the water quality of green roofs.

4.6.1. Temperature

Buffam et al. observed a positive correlation between temperature and nutrient leaching (N and P) from green roofs [100], potentially attributed to enhanced microbial mineralization and organic matter decomposition with increasing temperature, resulting in elevated N and P concentration in runoff water. However, contrary findings from other studies [101,102] suggest that temperature does not significantly influence net mineralization rates of N and P in soils. Akther et al. discovered a negative relationship between nutrient concentrations (N and P) on green roofs and substrate temperatures [103]. The results imply that as temperature rises, the nutrient concentration decreases within the system. This phenomenon may be due to plants thriving under higher temperatures with improved nutrient absorption capabilities, surpassing the effects of microbial mineralization and nutrient decomposition caused by elevated temperatures, ultimately leading to reduced nutrient leaching.

4.6.2. Humidity

The impact of humidity on the water quality of green roofs was also examined by Akther et al. from a humidity perspective [103]. It was observed that an increase in humidity led to an increase in nutrient leaching while metal element leaching decreased. These findings suggest that in high-humidity environments, accelerated organic matter decomposition results in greater leaching of nutrients (N and P), whereas metals are retained due to enhanced precipitation or adsorption.

4.6.3. Precipitation

Studies have indicated that green roofs exhibit lower concentrations of ${\mathrm{NO}}_3^{-}$-N in their runoff during large precipitation events, while TP and ${\mathrm{PO}}_4^{3-}$-P concentrations are higher [104]. However, a separate study suggests that the concentration of TP on green roofs is not affected by precipitation events [105]. Buffam et al. investigated the impact of temperature, humidity, and precipitation conditions on the concentrations of dissolved nutrients, alkaline cations, and metals in runoff from green roofs [100]. The findings demonstrate relatively low levels of various element concentrations in green roof runoff during significant precipitation events. This study also highlights that temperature has a more substantial influence on water quality compared to humidity and precipitation for green roof runoff. Nevertheless, research also indicates that the magnitude of precipitation events significantly affects the water quality of green roof runoff [106]. The results reveal an increase in TN, ${\mathrm{NO}}_3^{-}$-N, ${\mathrm{PO}}_4^{3-}$-P, ${\mathrm{SO}}_4^{2-}$, and DOC concentrations in outflow with rising rainfall amounts as well. During larger rainfall events, TN and TP removal rates display a negative correlation with rainfall intensity [107], suggesting that nutrient removal efficiency by green roofs weakens with increased rainfall intensity.

4.6.4. Climate Zone

The meteorological conditions in the vicinity of green roof systems typically encompass factors such as temperature, humidity, and rainfall. It is imperative to particularly focus on the operational status and water treatment efficacy of green roofs within specific climatic regions.

Guo et al. conducted experiments in the Mediterranean climate region, where they planted 11 types of plants on green roofs, including five herbaceous plants, three subshrubs, and three shrubs [108]. Their research was carried out under the hot and dry conditions of the local summer and confirmed that shrubs and subshrubs exhibited a higher survival rate compared to herbaceous plants. This disparity may be attributed to the morphological characteristics possessed by shrubs and subshrubs that enable them to adapt to arid environments, such as thicker wax layers and more efficient leaf-shedding mechanisms. The impact of the Mediterranean climate on green roof vegetation is primarily manifested through factors like high temperatures and dryness during summer, which determine plant survival and growth performance on green roofs while indirectly influencing water purification effects. Rocha et al., also conducting related experiments in the Mediterranean region, observed better water purification effects during autumn and winter compared to spring and summer [57]. This discrepancy can be attributed to rainfall being predominantly concentrated in autumn and winter within the Mediterranean climate, whereas spring and summer are relatively drier seasons. During periods of increased rainfall, green roofs can effectively absorb and filter rainwater, thereby reducing runoff volume. During autumn/winter, when soil moisture levels are higher, plant growth remains robust with an increased root density that aids in retaining water infiltration capacity, consequently enhancing water purification effects.

Akther et al. investigated the impact of cold semi-arid climates on the water purification effectiveness of green roofs [103]. The region experiences prolonged and frigid winters with repeated freeze-thaw cycles, which may influence the mineralization and leaching of nutrients in green roof substrates, consequently affecting their water purification performance. Moreover, due to limited precipitation in the area, snowmelt plays a significant role in runoff and significantly influences the chemical leaching behavior of green roofs; indeed, similar chemical leaching behaviors are observed during rainfall events and snowmelt events. During rainfall events, higher nutrient leaching rates occur in spring due to increased soil moisture content and elevated growth medium temperatures that favor nutrient mineralization processes. However, lower nutrient leaching rates are observed during the summer and autumn seasons. It is important to note that these nutrient leaching and mineralization processes not only impact the water purification effectiveness of green roofs themselves but also pose potential risks to downstream water bodies. Therefore, a comprehensive understanding and effective control of these nutrient leaching and mineralization processes are crucial for optimizing design.

The study conducted by Sultana et al. involved a comprehensive assessment of water quality in rainwater collected from green roofs under tropical climate conditions, with a primary focus on indicators such as dissolved oxygen (DO), pH value, electrical conductivity, and temperature [109]. The findings revealed that the green roof system exhibited excellent water quality performance in tropical climates, enabling direct utilization of untreated rainwater for toilet flushing and garden irrigation purposes. Furthermore, all samples maintained temperatures within the standard range, which indicates effective heat regulation by the green roof system.

Rey and his team conducted research in a neotropical mountain climate characterized by a bimodal precipitation pattern, with two wet seasons (March–April and October–November) and two dry seasons (January–February and July–August) throughout the year [64]. The study categorized rainfall events into three groups based on their depth and duration: ‘large events’ (longer duration or greater depth), ‘intermediate events’ (falling between large and small events in terms of duration and depth), and ‘small events’ (shorter duration or lesser depth). The findings revealed that intermediate-sized rainfall events were linked to longer preceding drought periods, indicating that the dry substrate has the capacity to absorb more water during such events after experiencing an extended period of drought, thereby preventing rapid water loss. This phenomenon potentially enhances rainwater purification on green roofs. Additionally, following the end of a drought period, nutrients within the substrate may have accumulated to certain levels. Intermediate-sized rainfall events could transport these nutrients downstream through runoff, which might impact water quality. However, if both plants and substrate can efficiently utilize these nutrients, it may reduce nutrient losses via runoff while improving water quality from green roof outflows.

The experiment conducted by Ferrans et al. took place in the same location as Rey et al., which is characterized by a typical subtropical highland climate [14]. Bogota’s rainfall pattern exhibits a bimodal distribution, with increased precipitation during the rainy season leading to higher runoff from green roofs, while reduced precipitation during the dry season impacts the water retention capacity of green roof systems. Seasonal variations were observed in pollutant concentrations within the outflow from the system, with elevated levels of BOD and TSS during the dry season and heightened levels of COD and total coliforms during the rainy season. These fluctuations may be attributed to rainfall patterns, temperature fluctuations, and plant growth activity.

4.7. Processing Objects

The management of rainwater runoff has been a primary focus in the investigation of green roof systems, with extensive research also conducted on harnessing these systems to enhance the water quality of rainwater. However, there is limited research on integrating green roofs into wastewater treatment processes. The utilization of conventional green roofs for wastewater treatment may yield suboptimal outcomes; nevertheless, wetland roofs provide an effective approach for treating and reusing such waste streams. By utilizing wastewater as an irrigation source for green roofs, irrigation costs can be minimized, and a secure and environmentally friendly solution for waste management can be attained. The current state of research on the treatment performance of green roof systems in wastewater treatment is presented in Table 3.

4.7.1. Greywater

Petreje et al. used recycled crushed building rubble containing a large proportion of brick as the substrate for wetland roofs to treat greywater [74]. The findings demonstrated that wetland roofs significantly mitigated the concentrations of total nitrogen and orthophosphate in the greywater. Ramprasad et al. employed the GROW (Green Roof-top Water Recycling System) wetland roof system for greywater treatment, incorporating eight different plant species into their design [110]. The GROW system exhibited remarkable efficacy in treating greywater with removal rates of 91.7% for TN, 83.6% for ${\mathrm{NO}}_3^{-}$-N, 87.9% for TP, 90.8% for BOD, 92.5% for COD, 91.6% for TSS, 93.4% for PG, 91.4% for FC, 88.9% for TMA, and 85.7% for SDS. Thomaidi et al. utilized green roofs to address greywater generated from buildings [49]. Among their systems, Atriplex halimus planted and filled with a layer of vermiculite at a depth of 20 cm achieved optimal treatment performance with removal rates reaching up to 93%, 93%, 91%, and 91% for TSS, turbidity, BOD, and COD, respectively.

4.7.2. Blackwater

The green roof exhibits not only a significant treatment effect on greywater but also effectively processes blackwater for potential reuse. Bui Xuan et al. (2014) developed a shallow subsurface flow wetland roof system planted with Melampodium paludosum to treat septic tank wastewater [94]. The findings demonstrated that the system achieved an average removal rate of 88–91% for TN, 77–78% for COD, and 72–78% for TP, thereby meeting local standards for water reuse and surface water discharge. Thi-Dieu-Hien et al. designed four distinct plant-based shallow wetland systems to treat septic tank wastewater, among which the average COD removal rates ranged from 61% to 79%, TN removal rates ranged from 54% to 81%, TP removal rates ranged from 62% to 83%, and suspended solids exhibited an average removal rate of 88 ± 3% [25]. Furthermore, the treated water complied with national standards for both discharge and reuse.

4.7.3. Other Types of Water

Green roofs are capable of effectively managing not only rainwater but also various types of wastewater, such as greywater and blackwater. A study conducted by researchers investigated the influence of seawater irrigation on green roofs planted with salt-tolerant plants [111]. The findings indicate that both partial and complete utilization of seawater for irrigation purposes are viable options. Particularly, when alternating between seawater and tap water for irrigation every four days, there is no negative impact on plant growth. Additionally, irrigating with seawater exclusively every four days still enables satisfactory plant development. This study presents novel insights for future irrigation water sources in green roofs, proposing the utilization of seawater in regions experiencing water scarcity. Simultaneously, considering the prevailing issue of severe water pollution, it is worthwhile to explore the potential use of green roofs for treating contaminated bodies of water such as lakes and seas. This approach not only conserves irrigation water but also exhibits a positive impact on water purification.

Table 3. Overview of main studies concerning wastewater treatment through green roofs.

Processing Objects	Influent (mg/L)		Effluent (mg/L)		References
	N and P Concentration	Organic Concentration	N and P Concentration	Organic Concentration
Greywater	TN: 10.1 ± 2.7 TP: 7.6 ± 2.4	COD: 226 ± 60 BOD: 132 ± 36	TN: 4.9 ± 2.7 TP: 3.9 ± 2.1	COD: 20–36 BOD: 20 ± 11	[49]
Greywater	-	COD: 226 ± 60 BOD5: 132 ± 36	-	COD: 25 ± 17 BOD5: 14 ± 10	[95]
Greywater	${\mathrm{NH}}_4^{+}$-N: 1.9–3.3 TN: 2.3–4.6 TP: 0.34–0.36	COD: 234–313 BOD5: 121–149	${\mathrm{NH}}_4^{+}$-N < 3 TN < 4 TP < 0.4	COD: 53.6 BOD5: 4.1	[27]
Greywater	TN: 16.3 (HLR1) TN: 16.7 (HLR2) TP: 1.2 (HLR1) TP: 2.6 (HLR2)	COD: 635.2 (HLR1) COD: 1115 (HLR2) BOD5: 393.3 (HLR1) BOD5: 407.7 (HLR2)	TN: 1.2 (HLR1) TN: 1.1 (HLR2) TP: 0.4 (HLR1) TP: 0.3 (HLR2)	COD: 69.1 (HLR1) COD: 61.3 (HLR2) BOD5: 10.6 (HLR1) BOD5: 5.9 (HLR2)	[112]
Greywater	${\mathrm{NH}}_4^{+}$-N: 10.28 –14.56 ${\mathrm{NO}}_3^{-}$-N: 12.32–7.84 TP: 2.934–3.84	COD: 216–320 BOD5: 68–120	${\mathrm{NH}}_4^{+}$-N: 0.67–0.95 ${\mathrm{NO}}_3^{-}$-N: 1.2–3.5 TP: 0.8–1.4	COD < 10BOD5 < 20	[110]
Greywater	${\mathrm{NH}}_4^{+}$-N: 1.2 ± 0.3 ${\mathrm{NO}}_3^{-}$-N: 1.6 ± 0.3 TP: 0.7 ± 0.1	COD: 81.9 ± 4.1 BOD5: 19.0 ± 0.9	No clear change	BOD5 < 10	[113]
Greywater	-	COD: 87 (low level) COD: 495 (high level) BOD5: 20 (low level) BOD5: 164 (high level)	-	COD: 19 (low level) COD: 159 (high level) BOD5: 2 (low level) BOD5: 80 (high level)	[114]
Blackwater	TP: 5.8 ± 0.6	COD: 176 ± 43	TP: 1.3–7	COD: 25–65	[94]
Blackwater	TKN: 42 ± 7 ${\mathrm{NH}}_4^{+}$-N: 38 ± 2 ${\mathrm{NO}}_3^{-}$-N: 0.5 ± 0.3 TP: 1.5 ± 0.7	COD: 108 ± 53	TN: 14 ± 3 TP: 0.4 ± 0.3	COD: 32 ± 26,	[25]
Blackwater	TKN: 42 ± 7 ${\mathrm{NH}}_4^{+}$-N: 38 ± 2 ${\mathrm{NO}}_3^{-}$-N: 0.5 ± 0.3 TP: 1.5 ± 0.7	COD: 108 ± 53	TN: 10 ± 4 (HLR1) TN: 10 ± 2 (HLR2) TP: 0.7 ± 0.3 (HLR1) TP: 0.4 ± 0.3 (HLR2)	COD: 29 ± 16 (HLR1) COD: 34 ± 23 (HLR2)	[26]

Table 3. Overview of main studies concerning wastewater treatment through green roofs.

Processing Objects	Influent (mg/L)		Effluent (mg/L)		References
	N and P Concentration	Organic Concentration	N and P Concentration	Organic Concentration
Greywater	TN: 10.1 ± 2.7 TP: 7.6 ± 2.4	COD: 226 ± 60 BOD: 132 ± 36	TN: 4.9 ± 2.7 TP: 3.9 ± 2.1	COD: 20–36 BOD: 20 ± 11	[49]
Greywater	-	COD: 226 ± 60 BOD5: 132 ± 36	-	COD: 25 ± 17 BOD5: 14 ± 10	[95]
Greywater	${\mathrm{NH}}_4^{+}$-N: 1.9–3.3 TN: 2.3–4.6 TP: 0.34–0.36	COD: 234–313 BOD5: 121–149	${\mathrm{NH}}_4^{+}$-N < 3 TN < 4 TP < 0.4	COD: 53.6 BOD5: 4.1	[27]
Greywater	TN: 16.3 (HLR1) TN: 16.7 (HLR2) TP: 1.2 (HLR1) TP: 2.6 (HLR2)	COD: 635.2 (HLR1) COD: 1115 (HLR2) BOD5: 393.3 (HLR1) BOD5: 407.7 (HLR2)	TN: 1.2 (HLR1) TN: 1.1 (HLR2) TP: 0.4 (HLR1) TP: 0.3 (HLR2)	COD: 69.1 (HLR1) COD: 61.3 (HLR2) BOD5: 10.6 (HLR1) BOD5: 5.9 (HLR2)	[112]
Greywater	${\mathrm{NH}}_4^{+}$-N: 10.28 –14.56 ${\mathrm{NO}}_3^{-}$-N: 12.32–7.84 TP: 2.934–3.84	COD: 216–320 BOD5: 68–120	${\mathrm{NH}}_4^{+}$-N: 0.67–0.95 ${\mathrm{NO}}_3^{-}$-N: 1.2–3.5 TP: 0.8–1.4	COD < 10BOD5 < 20	[110]
Greywater	${\mathrm{NH}}_4^{+}$-N: 1.2 ± 0.3 ${\mathrm{NO}}_3^{-}$-N: 1.6 ± 0.3 TP: 0.7 ± 0.1	COD: 81.9 ± 4.1 BOD5: 19.0 ± 0.9	No clear change	BOD5 < 10	[113]
Greywater	-	COD: 87 (low level) COD: 495 (high level) BOD5: 20 (low level) BOD5: 164 (high level)	-	COD: 19 (low level) COD: 159 (high level) BOD5: 2 (low level) BOD5: 80 (high level)	[114]
Blackwater	TP: 5.8 ± 0.6	COD: 176 ± 43	TP: 1.3–7	COD: 25–65	[94]
Blackwater	TKN: 42 ± 7 ${\mathrm{NH}}_4^{+}$-N: 38 ± 2 ${\mathrm{NO}}_3^{-}$-N: 0.5 ± 0.3 TP: 1.5 ± 0.7	COD: 108 ± 53	TN: 14 ± 3 TP: 0.4 ± 0.3	COD: 32 ± 26,	[25]
Blackwater	TKN: 42 ± 7 ${\mathrm{NH}}_4^{+}$-N: 38 ± 2 ${\mathrm{NO}}_3^{-}$-N: 0.5 ± 0.3 TP: 1.5 ± 0.7	COD: 108 ± 53	TN: 10 ± 4 (HLR1) TN: 10 ± 2 (HLR2) TP: 0.7 ± 0.3 (HLR1) TP: 0.4 ± 0.3 (HLR2)	COD: 29 ± 16 (HLR1) COD: 34 ± 23 (HLR2)	[26]

5. Conclusions

The selection of appropriate plant species is a crucial factor in establishing green roof systems, as plants have the potential to impact the functionality, stability, and long-term viability of such roofs. Plants play a significant role in maintaining the physical, chemical properties, and microbial conditions of green roof substrates and can effectively reduce nutrient concentrations in runoff post-planting. It is advisable to avoid deciduous plants due to their leaf litter decomposition, which may elevate nutrient levels in runoff. The limited range of plant species suitable for rooftop growth is primarily attributed to specific environmental constraints. When utilized for wastewater treatment purposes, selecting plants with well-developed root systems, large leaf areas, and high growth rates can yield favorable treatment outcomes. Additionally, considering interactions between different plant species and substrates is essential for optimizing treatment performance from green roofs. Substrates also hold critical importance in green roof system design; varying materials and proportions may result in nutrient leaching during rainfall runoff, which could classify green roofs as sources of pollution. Therefore, it is imperative to carefully select materials that do not contribute to pollution while designing substrate mixtures appropriately for optimal performance. Thicker substrates exhibit the ability to decrease pollutant concentrations in effluent while providing effective water purification effects. Furthermore, incorporating suitable amendments can enhance pollutant interception capacity by green roofs. Consequently, when designing compositionally or materially diverse green roof substrates based on site conditions and actual requirements, it becomes necessary to determine substrate proportions, thicknesses, and amendment application methods.

The slope of green roofs does not significantly affect the retention and removal efficiency of most pollutants. This may be due to the fact that the actual roof slopes are usually relatively gentle. Furthermore, during the design phase of green roofs, operational conditions such as hydraulic loading rate (HLR), hydraulic retention time (HRT), and water feeding patterns should be taken into consideration. With the prolonged operation of green roofs, there is a potential for improved pollutant removal efficiency. Enhanced treatment performance can be achieved through mature plant growth, stable substrate composition, and a diverse microbial community. When establishing long-term existing green roofs with purification and restoration functions, careful selection of suitable plants and substrates is crucial to avoid the accumulation of pollutants from components and materials used. Local climatic conditions, including temperature, humidity, and precipitation, also influence the performance of green roofs. When designing green roof systems, priority should be given to the environmental adaptability of the plants and the stability of the substrate to cope with seasonal changes, extreme temperatures, and precipitation conditions in different climate zones.

6. Prospective

In general, while some studies have considered green roofs as a potential source of pollutants, these findings can be attributed to substrate and plant factors. However, specific research has demonstrated that green roofs, particularly wetland green roofs, exhibit significant potential in the treatment of wastewater. It has been discovered that properly designed green roofs in urban areas can effectively treat both greywater and blackwater. Consequently, they serve as promising on-site rainwater treatment technology and an integral component of sustainable urban design and green building practices in the future.

The article critically examined green roofs and investigated their efficacy in nutrient removal, organic pollutant degradation, and heavy metal contamination mitigation, thereby showcasing the immense potential of green roofs in promoting sustainable water resource management and enhancing water quality purification in future urban areas. Drawing upon existing research findings, the following avenues for further development are suggested:

(1) Future studies could concentrate on selecting plant species with superior pollutant removal capabilities to enhance the efficiency of water quality purification by green roofs. Additionally, attention should be given to mixed planting strategies aimed at augmenting biodiversity within green roof systems.

(2) To address concerns regarding substrate composition and leaching of pollutants from materials, it is imperative to conduct research focused on developing novel materials characterized by reduced nutrient leaching rates so as to minimize pollution risks associated with green roofs.

(3) While most investigations have primarily centered around plants and substrates within green roof systems, insufficient emphasis has been placed on comprehending the impact of microorganisms on water quality purification processes.

(4) Green roofs have demonstrated effectiveness in treating rainwater, greywater, and blackwater; henceforth, additional research endeavors can be undertaken to explore their applicability for remediating other polluted bodies of water, such as lakes or seawater.

(5) Limited attention has been devoted to investigating the treatment efficacy of pathogens, antibiotics, and recalcitrant pharmaceuticals present in wastewater using green roof technologies. Therefore, future studies should prioritize strengthening research efforts in this area.