Urban Integration of Green Roofs: Current Challenges and Perspectives

: Green roofs (GRs) are a sustainable alternative to conventional roofs that provide multiple ecosystem services. Integrating GRs into urban areas is highly relevant considering the rapidly increasing built-up in cities. Therefore, this paper systematically and comprehensively reviews the recent literature from 2011 to 2019 on GRs to identify the challenges and perspectives related to the urban integration of GRs. The review suggests that the effectiveness of GRs in delivering ecosystem services is largely dependent on context-speciﬁc parameters such as weather conditions and existing construction or design-related parameters. Integrating GRs into urban areas can be challenging given the diversity of actors, functions, and conditions characterizing these areas. Although signiﬁcant research has already been conducted on GRs, research covering more geographical locations and contexts is needed. The review points out the need to include future urbanization scenarios, such as tall buildings while analyzing the impact of GRs on ecological networks. Additionally, the review emphasizes the inclusion of urban morphological parameters alongside an analysis of the impact of GRs on microclimate regulation and air quality. In terms of social acceptance, this review points out the need to consider the temporal cycles of vegetation for noting users’ perspectives. Additionally, further research is required on the social impact of GRs, considering their inﬂuence on property prices. Lastly, the review stresses the need for more city-scale studies on the impact of GRs on ecosystem services.


Introduction
Unprecedented urbanization in the world has given rise to enormous challenges associated with energy consumption, social inequality, air and water pollution, and resource depletion resulting in a massive strain on urban systems [1]. The United Nations (UN) has estimated that around one-third of the world's population will live in the cities by 2050 [2]. Thus, it is essential to significantly transform the way urban spaces are built and managed to ensure the sustainable development of cities [3].
At present, urban ecosystems consist of a large proportion of built infrastructure along with high population densities [4,5]. Urban ecosystems experience the urban heat island (UHI) effect, air and water pollution, and frequent flooding, owing to the lack of green spaces [6][7][8]. Impervious surfaces such as building roofs are abundant within urban ecosystems, occupying around 20-25% of urban surfaces. These are located on buildings that consume a significant amount of energy and are responsible for substantial greenhouse gas (GHG) emissions [9]. Given this background, nature-based solutions such as green roofs (GRs) are increasingly gaining popularity due to their positive effect on urban ecosystems [10][11][12].
GRs are defined as living vegetation planted on the roofs of buildings [13,14]. GRs can improve the energy performance of the buildings and aid in combating the UHI effect by reducing the atmospheric temperature as well as providing human thermal comfort [15]. the papers by screening the titles and abstracts and eliminated those articles that did not focus on understanding the impact of GRs on the selected aspects. Additionally, to achieve a feasible number of publications for review, we only focused on papers that compare GRs with bare, impervious, and conventional roofs. Papers including other kinds of green urban infrastructure were not reviewed.

Results
This section presents the number of papers selected for review (Section 3.1) and reviews them by discussing the contribution of GRs to the six defined aspects (Section 3.2).

Selection and Classification of Literature
The search strategy resulted in a total of 739 papers. We read the abstracts of all 739 papers and, based on the filtering process outlined in the previous section, selected the 158 most relevant studies for a full-text review. Figure 2 provides the details of the papers obtained initially, regarding each aspect, and the number of papers selected for the review based on the selection criteria. It is evident from Figure 2 that there is a significant amount of literature available on the contribution of GRs to water management and UHI mitigation as compared to the other aspects, based on our search strategy. Additionally, we classified the studies based on their scale; for instance, at a city, neighborhood, or building/prototype level ( Figure 3). Most studies were observed to have been carried out at a prototype or single-building scale. However, aspects such as water management, UHI, and feasibility consist of a few studies which were carried out at city and neighborhood scales. For the aspects of air quality and biodiversity, most studies were experimental and carried out on specific buildings or prototypes. In the case of social acceptance, the studies discuss the opinions of various stakeholders; therefore, the scale of the study was not relevant.

Biodiversity
Studies show that GRs support various types of flora, such as herbaceous perennials, succulents, indigenous species, wildflower species, and fauna such as birds, butterflies, and arthropods ( Table 2). Species such as ground beetles, collembolan communities, spiders, tree bugs, and hymenopterans are also found in abundance (Table 2). Mosquitoes are one of the species of arthropods, which can be harmful, as they are the carriers of severe diseases. Wong and Jim [36] reported that the mosquitoes' presence on GRs was less than in other green spaces; however, it was greater than on a bare roof. Other than arthropods, GRs have been observed to provide habitats for resident birds such as tree sparrows, spotted doves, and migrant birds [37][38][39][40]. These birds are attracted to GRs because of the availability of their diet, such as insects, grains, and berries [38,39]. With regards to flora, GRs provide space for conserving indigenous species. Sedums are often planted on GRs owing to their shallow root systems and high water-use efficiency, enabling them to withstand extreme weather conditions and the limited substrate depths of EGRs [41]. However, Vandegrift et al. [42] observed that some indigenous herbaceous perennials lasted as long on GRs as common sedum species. GRs also enable the enhancement of genetically diverse plant species that facilitate the spotting of birds [41]. Moreover, GRs support wild plants and flowers, which serve as a part of the food chain for essential pollinators such as wild bees along with enhancing aesthetic appearance.
IGRs are reported to provide a medium for conserving species [43]. Furthermore, it is worth noting that most of these species are observed in large numbers only on IGRs [44,45]. EGRs also have the potential to enhance biodiversity; however, this effect is lesser compared to IGRs [46]. Apart from this, the height of the building and size of the GR [40] also act as determining factors for the enhancement of flora and fauna. Additionally, Joimel et al. [47] argue that GRs can act as a medium for connecting various urban green areas, thereby providing the potential for habitat connectivity and maintenance of the ecological network.   Table 3 presents the water retention capacities given in the selected literature and the type of roof that was studied along with its respective details. Due to better feasibility, most studies focus on EGRs. We noticed that in some of the cases, GRs could retain the rainfall/runoff completely. In some of the studies, the retention capacity of a GR was reported to be below 10%. The parameters affecting the retention values are rainfall intensity and duration for most of the studies. The retention capacity of a GR is lower in the case of high-intensity rainfall in a short time as compared to medium intensity rainfall for a relatively long time. Rainfall characteristics have been discussed to be the most important parameter for deciding the water retention capacity of GRs. Additionally, evapotranspiration and the antecedent dry weather period (ADWP) also influence the retention capacity of GRs.  EGR, experimental study, rainfall intensity, duration, substrate depth. Retention less effective with heavy rainfall regime in the region; however, peak retention and delay was still significant. Substrate depth is also one of the important parameters [63,71]. Generally, deeper substrate corresponds with better retention capacity. However, a deep substrate might not be feasible for implementation. The evidence suggests that EGRs with shallow substrates are perfectly capable of managing stormwater. For instance, Wong and Jim [74] observed that both IGRs and EGRs have a significant performance. Some studies discuss various materials to be used in the substrate for the better retention of GRs as well [55,59,81]. The use of organic matter or recycled materials can be helpful [81]; however, this parameter has been of lower importance as compared to other parameters [59]. Vegetation type also influences the water retention capacity of GRs. Taller plants with deep roots can be useful for retaining water as compared to sedum species [81]. In addition to the aforementioned design parameters, field capacity, the capillary storage of substrate, and drainage routing also play an important role in water retention by GRs [57].

Flood Mitigation
This section discusses the six selected studies that were carried out at a larger scale (at a neighborhood or city level) to understand the capacity of GRs to reduce the impact of floods.
Versini et al. [85] reported that when more than 50% of potential roofs are green (EGRs), the reduction in hydrological response in terms of peak and volume of runoff can reach around 20% in the case of rainfall with a return period lower than ten years. However, the impact of EGRs for intense rainfalls is marginal. The performance of a GR depends upon the precipitation and initial substrate saturation. Higher precipitation results in higher substrate saturation. If the GR is experiencing frequent rainfalls, the discharge is higher with subsequent rainfalls depending on the substrate saturation level. Thus, Versini et al. [86] argue that GR performance is more effective in the case of short-duration (but sometimes intense) rainfall. Additionally, they observe that the total area of GRs is more important than their location in terms of reducing the total runoff volume. However, the spatial distribution of GRs also affects the reduction in peak discharge; for instance, greening the roofs in the upstream region is observed to be more beneficial in delaying the catchment runoff rather than downstream where cascading effects can occur [86]. Schmitter et al. [87] observed that the annual volume reductions were low but substantial in terms of flood protection when all traditional roofs were converted to GRs.
Mora-Melià et al. [88] reported that floods could be avoided if at least 50% of the building roofs in the study area were covered with GRs in the case of moderate rainfall. For extreme storm events, semi-intensive GRs covering 60-100% of the study area could prevent floods. The efficiency of GR implementation is suggested to be further dependent upon the vegetation type and substrate depth. However, even if IGRs are implemented at a city scale, Ercolani et al. [89] have observed that efficiency is better for smaller storm events. They also point out that the efficiency is dependent upon the sewer characteristics and structural improvements in terms of sewer relining or the enlargement of conduits. Additionally, EGR performance is observed to be marginally useful in the case of flash floods [90].

Runoff Water Quality
Urban stormwater runoff carries urban non-point source pollution, affecting water quality [91]. GRs have the potential to reduce runoff quantity, while their impact on runoff quality is debatable. We identified around 19 studies related to this. We mainly observed studies examining the concentrations of nutrients, namely total nitrogen (TN), total phosphorus (TP), total suspended solids (TSS), chemical oxygen demand (COD), ammonium compounds (NH 4 + -N), phosphates (PO 4 3− ), and nitrates (NO 3 − ) (Table 4). Moreover, the literature also shows evidence related to hard metals such as zinc (Zn), iron (Fe), copper (Cu), and lead (Pb) ( Table 5). In Tables 4 and 5, we report whether the GRs are the source or sink of the nutrients and metals, respectively. GRs are a source of pollutants if the pollutants are greater in the runoff than in the rainwater or in the runoff from bare roofs; otherwise, they are sinks [92].
When GR runoff leachate is compared to the initial rainwater quality, the concentration of TP, TN, K, nitrates, and phosphates are higher in GR runoff, meaning GRs are a source of these nutrients [93,94]. Nevertheless, when compared to bare roofs, GRs, especially EGRs, often act as sinks for nitrogen, nitrates, and TSS [95,96]. However, the concentrations of phosphorus and phosphates are usually similar to the runoff observed from bare roofs or are higher in GR runoff due to the presence of fertilizers in the substrates [97]. Thus, the review suggests that GRs are a significant source of phosphates. Although GRs are a source of phosphates in the runoff, some studies have reported the concentrations to be well within the standards for non-potable water use in some regions [98,99]. Therefore, there is a potential to reuse the runoff from EGRs for non-potable purposes and avoid the direct flow of runoff into water bodies. Roof gardens or IGRs, due to excess fertilizers and thicker substrates, release a significant amount of total phosphorus and potassium and so the runoff cannot be reused [10,[99][100][101].
Another finding is that the concentration of TSS, COD, and ammonium in the runoff decrease over the years, whereas the concentration of nitrates increases over time, and that of TP remains the same [93,102]. Usually, the first flush after an intermittent dry period results in many nutrients in the runoff. The concentrations reduce in consequent rainfalls [103]. Mostly, the studies suggest that GRs appear to be a source of nutrients such as nitrates and phosphates. However, amending the GRs with appropriate substrate materials, thickness, and vegetation types could improve the runoff water quality. •-source; x-sink; o-source (acceptable water quality); ∆-source (values greater than the water quality standard).
Apart from nutrients, runoff from GRs also contains metal pollutants. Table 5 suggests that, in the reported studies, GRs behave as a sink for most metals. However, aged GRs (43 years old), although aiding in reducing the concentrations of nutrients, release additional amounts of metals such as Cu, Zn, Fe, and Pb in the runoff as compared to bare roofs [94]. Supporting this argument, Okita et al. [107] also reported that a sixyear-old GR has similar concentrations of Cu as compared to bare roofs. Apart from this, Alsup et al. [109] have observed that a GR's behavior as a sink or source of metals fluctuates depending upon the substrate properties and the dry and wet deposition of metals. Substrate materials such as expanded clay and organic materials have been effective in reducing Zn and Cu concentrations [110]. •-source; x-sink; o-source (acceptable water quality); ∆-a source (values greater than the water quality standard).

UHI
In this section, we report the results of 39 studies and discuss the impact of GRs on the UHI effect regarding the following four aspects: roof surface temperature, energy savings, air (outdoor, area above the roof) temperature, and human thermal comfort (indoor and street-level temperatures).

Roof Surface Temperature
The reduction in surface temperature due to the implementation of GRs ranges from 1 to 46 • C depending upon various factors, such as the type of GR, weather conditions, climate zones, substrate properties, and type of vegetation ( Table 6). The surface temperature after roof greening does not reduce significantly in winters and cold climatic conditions. Another important finding is that GRs have different impacts on roof surface temperature depending upon the time of the day. The direction of heat flux is top to bottom during the day and reverses during the night [111]. As a result, the GR surface cools during the day and warms at night [112,113]. This effect is more prominent in IGRs than EGRs as deeper systems gain more heat [112].  Lin et al. [113] Night-time: Afternoon: 1.0 • C to 13.6 • C-Taipei 6.5 • C to 17 Vegetation properties such as high foliage density and high leaf area index (LAI) aid in improving the performance of GRs [114,115]. Additionally, lower albedo and evapotranspiration due to large soil water content in GRs also aid in lowering surface temperatures [116]. However, Dvorak and Volder [117] reported that GRs with low soil water content are also equally effective in reducing the surface temperature in extreme dry summers. Table 7 presents the findings from the selected literature and indicates that GRs help with saving considerable energy, especially in summer. In summer, a reduction in the heat gain from the roof slab leads to a decrease in the cooling loads. However, in winter, heating loads have been reported to increase due to a reduction in heat gain during the day and an increase in heat loss during the night [112].

Energy Savings
Berardi [128] reported that using appropriate substrate materials (greater depth) and vegetation types (high LAI) with lower U-values (thermal transmittance) can aid in reducing the heating load in winter. Karachaliou et al. [115] have observed that IGRs are effective in reducing heating loads, but not as effective at reducing cooling loads in summer. The effect of the reduction in cooling and heating loads is more prominent on the topmost floors, and the rooms on other floors only experience marginal changes [115]. Apart from this, additional insulation has also been observed to aid in reducing heating loads; however, the efficiency is dependent upon climatic conditions. For instance, Kotsiris et al. [129] observed that insulation is ineffective in the case of the Greek Mediterranean climate, whereas Gagliano et al. [114] and Ðordević et al. [130] observed that insulation works best for heating loads in a moderate continental climate. Table 7. Energy savings due to GR implementation, as per the selected literature.
China, large-scale EGR and IGR, humid subtropical climate.
Serbia, EGR, cold climate, additional thermal insulation and substrate thickness.
Foustalieraki et al. [123] Reduction in the annual cooling load up to 18.7% and in the heating load up to 11.4% in total.
Athens, Greece, EGR, cold period, dense foliage plants resulted in lower temperatures.
Athens, Greece, IGR, summer period, Mediterranean climate. Plants with low absorptivity to solar radiation, dense foliage resulted in lower temperatures.
Berardi [128] Building energy demand reduced by 3%. Heating saving by 9%, cooling saving by 4% Toronto, Canada, EGR, semi-continental climate. Increasing the soil depth is more important than increasing the LAI.
Energy savings were related to the LAI in summer and to the soil depth in winter.
Ujjain, India, rooftop garden, temperate climate, summer day. Thicker soil, tall foliage reduced the heat flux.
Outside Air Temperature Table 8 shows the reductions in air temperature near the roofs after the installation of GRs, as reported in the selected literature. Experimental studies have measured the impact of GRs on air temperature at a height ranging from 10 cm to 300 cm above the roof. The temperature reduction ranges from 0.5 to 6.1 • C. However, it is essential to know that, except for J. Park et al. [132], other studies reported a reduction of only up to 2 • C. Amongst the studies that used a model-based approach, Heidarinejad and Esmaili [133] reported the highest reduction in air temperature (13 • C). The reason could be the consideration of plant metabolism, including photosynthesis in the model. Validation showed that the model results were similar to a real situation.
It should be noted that the temperature reduction is higher at night as compared to daytime due to the accumulation of thermal energy-which is highest in bare roofs relative to GRs-rather than the latent cooling of GRs [134]. The cooling observed over the GR usually corresponds to the cooling on the surface. If the surface temperature of well-watered and dry GRs was almost the same [134,135], the air temperature above them varied. The main difference the review suggests between these studies is the difference in plant species and substrate depth that reduce the overlaying air temperature, possibly due to higher evapotranspiration. Sedum-type vegetation and thinner substrates might not be an appropriate choice if the aim is to reduce the daytime temperature [135].
Altogether, the evidence suggests that GR installation results in the cooling of the air temperature near the roofs. However, in extreme summers, suitable plants and substrates with appropriate water content might be needed to enhance GR performance.

Human Thermal Comfort
In terms of human thermal comfort, we selected studies showing indoor temperature reduction and a reduction in temperature at a pedestrian or street level. Table 9 shows that GRs reduce indoor temperatures up to 4.2 • C depending on the diurnal weather conditions. The cooling is more significant during the day than at night. In some instances, at night, an increase in indoor temperature was observed due to the insulation provided by the GR [116,143]. This could be easily be solved by improving the natural ventilation in the building [114]. Additionally, selecting appropriate substrates and species other than sedum can also enhance the results. Temperature reductions at the pedestrian level are not as pronounced as those observed at roof level. The temperature reductions range from 0.1 to 1.7 • C, mostly depending upon the type of GR (Table 10). Compared to EGRs, IGRs provide better pedestrian thermal comfort. The review suggests that the installation of GRs on low-rise buildings has a better impact on the reduction in temperature at the pedestrian level (see Table 10).
Jadaa et al. [145] reported a reduction of 12-14 • C in mean radiant temperature (MRT) on a hot summer day, and around 1-4 • C on a winter day after retrofitting GRs in a small neighborhood with low-to medium-rise buildings in the UAE. Installing an IGR is reported to reduce the physiological equivalent temperature (PET) by 10.9 • C on a hot summer day [120]. The installation of an IGR on a building reportedly resulted in a reduction in the UTCI of 5.5 • C [120].
Altogether, installing GRs improves human thermal comfort, though there can be slight warming observed during the night. Furthermore, there is only a limited impact of GRs on pedestrian thermal comfort, but this can be more significant in the case of low-rise buildings. Table 9. Indoor temperature reduction after implementing GRs, as reported in the selected studies.

Air Quality
We discuss the results from nine studies on GRs in the context of air quality. The studies primarily discuss the effect of GRs on the concentration of air pollutants such as O 3 (ozone), CO 2 (carbon dioxide), SO 2 (sulfur dioxide), NO 2 (nitrogen dioxide), NO (nitrogen oxide), and PM 10 and PM 2.5 particulate matters. Table 11 provides the details of the studies that focused on the effects of GRs on particular pollutants. Based on the observations listed in Table 11, the installation of GRs leads to reducing the concentrations of air pollutants, depending upon various conditions. For instance, Moghbel and Erfanian Salim [147] observed CO 2 absorption to be the highest during the early morning and late evening and the lowest during peak traffic hours. Similarly, during morning rush hours, PM 10 and SO 2 were also reported to be higher than the WHO standards [151]. Apart from this, Agra et al. [148] reported that the GRs with Sedum Sediforne added to the high CO 2 concentrations in the city by emitting CO 2 . Regarding PM 2.5 , concentrations near the roof were reported to be reduced as compared to at street level. This was mainly attributed to less vertical mixing and high wind shear induced by higher thermal stability near the GR [149].
Concerning O 3 reduction, large, continuous GRs on taller buildings perform better in reducing O 3 near the rooftop [152]. Two more studies [153,154] analyzed O 3 reduction within a street canyon. Park et al. [154] considered a street canyon with an aspect ratio (AR, height-to-width ratio, H/W) of 1.0. The study considered NOx-O 3 reactions and observed that O 3 concentrations were dependent upon these reactions. For instance, O 3 is low in the areas where NOx is high. Primary pollutants such as NOx (nitrogen oxides), VOCs (volatile organic compounds), and CO (carbon monoxide) were concentrated near the street, whereas O 3 was found more in the region above the buildings. Roof cooling increased the in-canopy winds, which facilitated the incoming flow of cleaner air (with O 3 and without primary pollutants) and the outgoing flow of more polluted air (with primary pollutants).
When the aspect ratio was varied along with the leaf area density (LAD) in a similar setup, Moradpour et al. [153] suggested that for lower aspect ratios, the concentrations of ozone are higher as compared to H/W = 2.0. The difference can be attributed to the structure of the flow, where H/W = 2.0 contained two vertically aligned vortices, and scenarios with H/W ≤ 1 contained one clockwise vortex. Thus, the authors suggested that the GRs are more useful for deep street canyons.
Similar to the previous studies, Baik et al. [155] also assessed the impact of GRs in an urban street canyon; however, the pollutants were considered to be non-reactive. A similar result was obtained where the cooling from the GRs flowed into the street canyon, resulting in a strengthened canyon flow, which enhanced the pollutant dispersion near the road, thereby decreasing the pollutant concentration and improving air quality at a pedestrian level. However, the authors pointed out that the degree of air quality improvement near the roads depended upon the ambient wind direction. This phenomenon stands true in the case of Moradpour et al. [153] and Park et al. [154] as well, especially in the case of NOx concentrations as their movement is initiated due to the cooling offered by GRs. One should note that these three studies incorporated primary urban morphologies with simplified street canyon aspect ratios. Moreover, as observed in the study by Tong et al. [149], where an existing GR was used, the pollutant concentrations at street level did not reduce as much. Moreover, in reality, the aspect ratios in urban environments are complex, and buildings in a canyon are of different heights. Altogether, the evidence suggests that GRs have the potential to reduce toxic pollutants when implemented at a large scale, suggesting a positive impact on air quality.

Social Preference and Acceptance
We found around 17 studies discussing the opinions of various stakeholders, such as architects, engineers, property developers, students, and common people, who are potentially the ones to implement and utilize GRs. Figure 4 represents the benefits perceived by stakeholders such as residents and professionals such as architects, engineers, and technical experts. Two studies considered in the review discussed the perception of residents, whereas five studies were observed to discuss the perception of professionals in the field.
Despite the perceived benefits, there are also some factors hindering the acceptance of GRs among stakeholders ( Figure 5). Most studies suggest the initial construction costs, maintenance costs, and lack of awareness to be the main hindrances for the widespread acceptance of GRs. Moreover, the lack of knowledge and skilled labor for implementation and the lack of government incentives are additional concerns.
Acceptance of GRs among common people is related to their socio-economic backgrounds such as age, income, education level, attitude towards environmental issues, and social norms. For instance, L. Zhang et al. [161] observed that people with higher education and income and with an interest in environmental issues favor GRs. Fernandez-Cañero et al. [162] also identified age as a parameter, wherein younger people proactively support the development of GRs in cities. Social norms-for instance, support from social networks, i.e., amongst family and friends, for investment in green initiatives-were also identified as one of the parameters for considering GR adoption. Furthermore, a perceived difficulty in being environmentally advanced, i.e., the non-availability of sufficient time and resources to adopt GRs, also plays a vital role in the adoption of GRs. Thus, for the adoption of GRs, it is vital to not only increase awareness but also to reduce the costs associated with GRs. Apart from this, Fernandez-Cañero et al. [162] have also pointed out the importance of the type of housing where people reside for the adoption of GRs. For instance, it is easier to install a GR on a privately owned house rather than on a rented apartment.  [156] and Kim et al. [157]; perceptions of architects, urban planners, and others reported by Tabatabaee et al. [158], Sangkakool et al. [11], Tam et al. [159], and Rezvani et al. [160]).
The visual appearance and aesthetics of GRs play an essential role in their adoption by urbanites [163][164][165]. Perceptions based on experience and expectations and accessibility also play a crucial role in the demand for and favor of GRs. Loder [166] observed that the participants' previous experiences of nature and of narratives about native habitats mediated their perceptions of GRs in a central business district. For instance, prairiestyle vegetated GRs are found beautiful by the people of Chicago due to their previous experiences with areas outside of Chicago. In contrast, people in Toronto view real nature as forests and lakes. Thus, although they considered prairie-style GRs beautiful, they did not relate them to nature. Loder [166] also reported that the expectation of daily life experience plays an important role. The GR space could be optimized to mimic the memories of natural experiences, depending upon the visual or physical access demands as per the employees in the business districts. Along this line, Mesimäki et al. [167] observed that GRs provide experiential and recreational benefits. The authors' findings indicated that even a small but aesthetic accessible GR has the potential to offer a moment of respite in the middle of strenuous urban life [167].
Our review suggests that society prefers GRs, and there is the potential for the social acceptance of GRs if they have an affordable and aesthetically pleasing design. Additionally, user-specific experiences and expectations are essential to consider. Furthermore, there is a need for awareness programs related to the benefits of GRs to increase their outreach.

Feasibility of GRs
In this section, we discuss the findings from our review related to the feasibility of GRs, in terms of financial, economic, social, and environmental aspects. We short-listed ten studies that discuss the feasibility of GRs with the inclusion of several costs and benefits.
We also found three other studies discussing GRs in terms of private and public costs and benefits. Table 13 presents the types of costs and benefits considered in each of the studies. In the table, the review suggests that most studies consider only a few aspects of feasibility; however, some studies consider a wide range of costs and benefits which might increase their relevance and greatly influence the conclusions drawn. Table 12 presents the lifespan, scale, and city studied along with the key results in terms of the method chosen for the feasibility analysis. There are different methods used to study feasibility, for instance, benefit-cost ratio (BCR), net present value (NPV), life cycle costing (LCC), and payback period (PBP). Among the selected studies, Teotónio et al. [173] incorporated the most types of costs and benefits in their analysis. This study was done only for commercial and residential buildings in Lisbon at financial, economic, and socio-environmental levels. Socioenvironmental benefits were the highest, followed by economic benefits, which included the increase in property value. The financial NPV was negative, given the high installation and maintenance costs. Considering all levels, the benefits were greater than the costs. The study also suggested that IGRs, though they have high costs, have more benefits than EGRs, one of which is the increase in recreational value.
Other studies have considered comparatively fewer benefits than Teotónio et al. [173]. Most of the studies have a BCR greater than one and an NPV greater than zero. However, Ziogou et al. [174], Ziogou et al. [175], and Shin and Kim [176] reported that the benefits of GRs are almost equal to or less than the costs. The primary reason for this is the initial installation cost of a GR. The cost of GRs varies depending upon the municipalities and the types of costs considered. One possible reason could also be the consideration of limited benefits. It should be noted that it is complex to measure all the benefits of GRs in monetary terms. If only the widespread benefits such as energy savings, emission reductions, and stormwater management were considered, the feasibility of GRs would be debatable. However, when more benefits are considered in the analysis, GRs appear to be more feasible.
Until now, we classified the benefits in terms of social, financial, economic, and environmental aspects. However, analysts often consider the benefits in terms of private and public as this clarifies the concerns of private developers and the government [27]. The private benefits, availed by private investors, cover the increased life span of the roof cover, the reduction in seasonal energy consumption, the added aesthetics of GRs, noise reduction, and the increase in property value. At times, additional subsidies provided for the adoption of GRs are also categorized as private benefits [177]. The public or social benefits include improvement in the UHI effect, the betterment of stormwater quality, reduced GHG emissions, and enhanced biodiversity [178]. Although the overall benefits of GRs exceed the costs, it is essential to understand if there are benefits at a private level, as this is what decides the adoption of GRs. Claus and Rousseau [27] observed that without government subsidies, the costs are higher than the benefits at the private level. However, providing subsidies might be costlier to the municipality if GRs are not offering enough benefits at a city scale. Mullen et al. [178] observed that the subsidies reduced the NPV at the public level. The increase in the property value can increase the private benefits more than the costs even without the provision of subsidies [177].
Overall, the evidence suggests that the implementation of GRs is feasible in most aspects, even at private and public levels, considering all the benefits of GRs in the analysis. However, more studies are needed that incorporate the benefits of GRs in a comprehensive manner and at different levels, which elaborate the benefits for the involved stakeholders.

Quantity of Available Literature
The paper reviewed a total of 158 papers, which were identified using a systematic search strategy. Based on the selection criteria, the documentation (number of research articles) on GRs' contribution to each ecosystem service varied significantly (Figure 2). According to our review, the GRs' contributions to water management (61 studies) and UHI (39 studies) were more widely studied than the other aspects. The aspect of GRs' contribution to the improvement of air quality appears to be limited. Apart from this, most studies selected in this review were carried out at a building level. There were a few studies that were done at a city and neighborhood level in terms of water management, UHI, and feasibility. However, more studies at a larger scale for other aspects such as biodiversity and air quality can boost motivation for the integration of GRs into urban areas.

Current Status and Challenges
This section discusses the reviewed evidence and identifies the challenges of incorporating GRs into urban areas.
Regarding the impact on biodiversity, the review suggests that GRs on buildings with lower heights have been observed to support species in a better way [50]. Given that urban areas are densifying and becoming more compact, cities are predicted to grow vertically. Thus, the height of buildings is an essential parameter for analyzing GRs' impact on biodiversity. In the review, we did not find enough studies in the literature on the effect of height on GRs' impact on biodiversity. We found only two studies where Braaker et al. [50] considered buildings of 15 m or less and Wang et al. [40] considered a 50 m tall roof garden. We need more studies analyzing the impact of building height on GR biodiversity.
Similarly, there is a paucity in the literature on GRs' role in ecological connectivity. We found that only Braaker et al. [50] and Joimel et al. [47] considered this aspect to any extent. As GRs are considered to be a substitute for green areas on the ground, they are supposed to facilitate the movement of organisms through urban landscapes [183]. Therefore, more evidence is needed to understand the role of GRs in strengthening ecological networks at an agglomeration scale. Apart from this, our review also pointed out a concern in terms of mosquito abundance on GRs [36], which can have an effect on the health of people, especially in dense urban areas.
Our review suggests that GRs aid in water management, owing to their high water retention potential. However, the retention is high only for low-to moderate-intensity rainfall. Furthermore, GRs can be one of the factors in avoiding flood-like conditions. Still, other solutions might be needed in case of extreme rainfall. The review also suggested that the runoff from GRs also acts as a source of phosphorus in the case of most commercial substrates, which can pollute water resources. The concentration of pollutants in EGRs is lower than in IGRs due to the existence of fertilizers in IGRs. Alternative substrates and vegetation types were also observed to reduce the concentrations of nutrients or metals.
Regarding UHI mitigation, our review suggests that GRs have a limited impact on pedestrian thermal comfort. A significant impact can only be found with low-rise buildings. Building height has an influence on the impact of GRs on the temperature at street level. Besides, urban morphology is observed to play an important role in temperature regulation [184]. The performance of GRs may also vary depending upon the urban morphology of a block and its surroundings. This should be further analyzed in future studies.
Apart from this, our review suggests that GRs substantially reduce roof surface temperatures compared to bare roofs. Additionally, we observed that most studies reported significant reductions in energy demand and consumption due to a decrease in cooling loads. Heating load reductions were also observed in cases where appropriate substrates, thermal insulation, and plant species were used. However, these effects are prominent only on the top floors. GRs also decrease the indoor temperature of a building. However, slight warming can be experienced during the night. It should be noted that the maintenance of GRs is important for the desired benefits, especially in extreme weather.
Regarding the impact on air quality, our review suggests that GRs reduce the concentrations of pollutants in the air near the GR. The literature on the improvement in the air quality near the street also shows a positive impact on air quality. However, the urban morphology used in the models for air quality was quite simplified, and the results might vary in reality due to more complex morphologies.
In terms of social preference and acceptance, our review suggests that GRs with lowcost and high aesthetic value are in demand. Moreover, our review suggests that hindrances such as high costs and maintenance and lack of knowledge often affect GRs' acceptability in urban areas. In the feasibility studies, high installation costs were considered as a barrier at the private level, suggesting the need for subsidies. Apart from this, the adoption of GRs also depends on the type of housing. For instance, the UHI benefits such as energy savings are mainly reported to be significant at top floors and negligible elsewhere. Ergo, it can be challenging to get the approval of all residents of a multi-story building, given that the benefits are unevenly distributed.
Concerning social preference, most studies considered well-maintained GRs, which are effectively green, for understanding users' perspectives. However, due to extreme weather conditions in some regions, the vegetation during the months of winter or summer can be brown and dry, resulting in a different opinion of users. More extensive studies are required to study the social acceptability of GRs during the dry months when vegetation is dried.
In the review, we observed that the social benefits in the feasibility analyses were only considered in terms of public health, but the impact of GRs on social cohesion/inclusion was not considered. Greening strategies, at times, tend to increase property prices and result in gentrification and the marginalization of the poor [185]. More studies are needed to understand the social impact of GRs at the neighborhood scale.

Future Perspectives
This section discusses the future perspectives for research into the urban integration of GRs. The future perspectives in terms of further research are as follows:

•
Regarding biodiversity, future studies should consider taller buildings while analyzing GRs' impact on urban biodiversity. Moreover, the studies analyzing the role of GRs in ecological connectivity need to consider the height of buildings. Future research should also focus on mainstreaming GRs in a city-scale habitat network analysis to connect existing ecological areas.

•
Regarding water management, as runoff water quality from GRs is a concern, according to our review, further studies should focus on how the runoff from GRs can affect surface water sources and urban water management.

•
In terms of UHI mitigation, as the urban morphology of a neighborhood can be a contributing factor in the impact of GRs on pedestrian thermal comfort, future research should include urban morphological parameters in the study to understand the effect of GRs on pedestrian thermal comfort in a better way.

•
Concerning air quality, further research should involve more realistic morphologies while analyzing the impact of GRs on air quality. Additionally, more studies at a large scale would enhance the level of evidence for the impact of GRs on air quality.

•
Regarding social acceptability, low-cost GRs are essential for better integration into urban areas. Nevertheless, the aesthetic value of GRs also can boost their implementation. To understand the acceptability of GRs, further research should focus on considering the temporal cycles on vegetation while observing users' perspectives.

•
Regarding feasibility, the cost of GRs is one of the concerns. While subsidies may be one of the solutions, research for identifying the affordable components of GRs can pave the way for large-scale adaptability.

•
Further research should focus on GRs' social impact, especially on property prices and the subsequent possibility of gentrification.

Conclusions
This paper systematically and comprehensively reviewed the recent literature on GRs to identify the challenges and perspectives related to the urban integration of green roofs. Our review suggests that GRs have the potential to deliver multiple ecosystem services at the same time. GRs are the most useful tool considering the rapid increment in the built-up and lowered availability of open spaces for intensive green infrastructure. However, the effectiveness of GRs in delivering ecosystem services is largely dependent on context-specific parameters, such as weather conditions and existing construction or design-related parameters. Integrating GRs into urban areas can be challenging given the diversity of actors, functions, and conditions characterizing these areas. Although significant research has already been done on GRs, there is a need for more evidence covering more geographical locations and contexts. Firstly, the review pointed out the need for including future urbanization scenarios such as tall buildings while analyzing the impact of GRs on ecological networks. Secondly, the quality of runoff from GRs remains a concern. Further research is needed on how this runoff can affect surface water sources and urban water management. Regarding UHI mitigation, the review emphasized the inclusion of urban morphological parameters alongside future analyses of the impact of GRs, as morphology plays a vital role in microclimate regulation. Similarly, considering realistic urban morphologies is necessary for analyzing the impact of GRs on air quality. In terms of social acceptance, our review points out the need to consider the temporal cycles of vegetation while noting the perspective of users. Additionally, further research on the social impact of GRs is needed, considering their influence on property prices. Lastly, the review stresses the need for more city-scale studies on the impact of GRs on ecosystem services, especially in terms of biodiversity and air quality.

Conflicts of Interest:
The authors declare no conflict of interest.