Review of Sponge City implementation in China: performance and policy

Urban flooding is one of the main problems faced by major cities around the world. Rapid urbanization and economic development are transforming land cover to become increasingly impermeable (Zhang et al. 2019; Sun et al. 2020) leading to more frequent flood hazards and combined sewer overflows (CSOs) when the amount of runoff exceeds the capacity of systems traditional drainage. The construction and expansion of cities evolve demolitions, vacancies, and redevelopment, each of which affects the movement of water on land surfaces by modifying the hydrological properties of the soil, land cover, and topography (Kelleher et al. 2020) and even modifying the original path of the water in some cases. According to the United Nations estimation, 70% of the world's population will live in urban areas by 2050 (Lund et al. 2019), resulting in increased urban flooding (Ren et al. 2020) due to increased peak discharge and volume (Elliott & Trowsdale 2007; Ahiablame & Engel 2012; Raei et al. 2019; Nowogoński 2020), deterioration of downstream water quality and reduced water surfaces (Kelleher et al. 2020). Besides, extreme meteorological phenomena are favored by climate change (Li et al. 2018b).

While increasing soil infiltration is a straightforward solution to mitigate surface runoff, its application demands engagement from different authorities and professionals, because it is affected by various complicated phenomena such as rainfall intensity, rain event duration, soil type, initial coefficient of saturation, slope, and land cover (Ren et al. 2020). For that purpose, alternative techniques of eliminating surface runoff in a more useful and environmental friendly manner, such as best management practices (BMPs) and low-impact development (LID), are being developed around the world (Raei et al. 2019), in the US, water-sensitive urban design (WSUD) in New Zealand and Australia (Radcliffe 2019), sustainable drainage system (SuDS) in the UK (Csicsaiova et al. 2020), and Sponge City in China (Jia et al. 2017; Nguyen et al. 2019; Zhang et al. 2019). Practices such as bioretention or rain garden (RG) (Wang et al. 2019; Jiang et al. 2020), rain barrels (RBs) (Jiang et al. 2020; Nowogoński 2020), grass swales (GS) (Davis et al. 2012; Mei et al. 2018; Gong et al. 2019), green roofs (GRs) (Liu et al. 2019b), ponds (Xu et al. 2018; Nahar et al. 2019; Wang & Guo 2019; Csicsaiova et al. 2020), and pervious pavement (PP) (Li et al. 2013; Ahiablame & Shakya 2016; Liu et al. 2020; Randall et al. 2020) are used in this concept to maintain the natural hydrologic cycle (Nahar et al. 2019; Csicsaiova et al. 2020; Sun et al. 2020).

Cities now house more than 59% of China's overall population (NBoSo 2017). Since 2013, 30 pilot cities have been selected to establish Sponge City initiatives in order to mitigate the effects of urbanization on the urban environment (Shao et al. 2020). Similar to the LID concept, the ‘Sponge City’ concept involves planning and engineering design, using green-grey drainage infrastructures, which, when combined, can mimic natural hydrologic processes and promote infiltration and bioretention processes to manage and control stormwater runoff, maintain groundwater recharge, and protect water quality downstream (Chan et al. 2018). The major goal of the first batch of 16 pilot cities, announced in 2015, was to retain at least 70% of annual urban rainwater as a control measure for small- to medium-intensity storms. The second batch of 14 pilot cities, announced in 2016, increased the aims to focus more on natural ecologic development, stormwater reuse, increasing water quality, minimizing waterlogging, and mitigating climate change impacts such as urban micro-climate. Sponge City, LID, and other alternative stormwater management systems cannot be treated as a general model that adapts worldwide, as they depend on the physiological characteristics of the catchment in question, region climate, and hydraulic and hydrological parameters. Urban flooding, stormwater management, CSOs, and water quality, are the primary focus of green stormwater infrastructure concepts (Dickin et al. 2020). There are over 30 different types of stormwater management technologies that, individually or in combination, promote the following six goals: infiltration, detention, retention, purification, harvesting, and drainage (Zhang et al. 2019). There is also an undeniable link between sanitation, health, and climate change, including modifications of waterborne, increases in antimicrobial resistance, and implications for mental health and well-being, such as increased stress, potential exposure to violence, and anxiety from lack of access to toilet facilities (Dickin et al. 2020). Lee & Bang (2000) demonstrated how Sponge City technology can improve water quality through runoff mitigation; however, sponge cities have been shown to be less efficient in regulating runoff in large storms. According to Griffiths et al. (2020), 19 out of the 30 pilot cities experienced flooding after their implementation. In this review, we look into the prevalent causes of these issues, in order to improve the long-term performance of sponge cities.

A wide range of research studies on the topic have focused on evaluating the effectiveness of Sponge City projects. Much prior research evaluated the effectiveness of sponge cities (Eckart et al. 2017; Leng et al. 2020; Nguyen et al. 2020), mostly based on stormwater runoff control, peak flow mitigation, and stormwater pollution removal rates. Sponge City has demonstrated good efficacy in managing both the quantity and quality of stormwater discharge (Li et al. 2019). Other research has included additional factors to the assessment. For example, Leng et al. (2020) have incorporated receiving water bodies as an indicator in their assessment. Alternatively, Kim et al. (2019) in their study, considered the impact of groundwater on infiltration and storage rates. Meanwhile, many studies have been conducted to assess the performance at different surface scale levels: a catchment scale (Ahiablame & Shakya 2016; Bae & Lee 2019; Randall et al. 2019), city scale (Liu et al. 2021b), pilot scale (Liu et al. 2020), and block scale (Palla et al. 2017). Furthermore, some other studies have been interested in the performance at a facility scale, such as Jia et al. (2016) studied the performance of RGs, Davis et al. (2012) studied the performance of GS, and Yu et al. (2021) focused their study on the PP. Moreover, as the focus of our analysis, LID facilities can also be evaluated based on their overall planning, working processes, and ecological impact.

Simultaneously, tremendous attention was dedicated to studying the shortcomings of the SPC concept, to evaluate the challenges and barriers in China that may stand against successful implementation of SPC projects (Zheng et al. 2016; Xia et al. 2017; Li et al. 2020). Systemic challenges related to the materials used, facilities planning and design, or the implementation design as a whole, such as privileging PP over other relatively efficient facilities, in the design due to its affordable price (Li et al. 2013) or due to a lack of vacant parcels for other LID facilities (Oates et al. 2020). At the same time, PP is particularly effective in mitigating stormwater runoff (Rodríguez-Rojas et al. 2018). However, the overuse of infiltration-based facilities can lead to substantial contamination of subsurface water allowing the runoff pollutants to reach the underground water (Yawen et al. 2020). Another study by Starzec et al. (2020) has shown that using infiltration-based facilities on land with a high groundwater table or compacted soil might result in a significant drop in infiltration rates, which directly increases in flood risks. Furthermore, the experiment of Liu et al. (2019b) is another projection of SPC implementation challenges, the study showed that if GRs are constructed with improper materials (substrate and plants), they can have a reverse effect and increase pollutant concentration rather than promote pollutant removal. Manifestly, financial limitations are the source of most of these challenges. These themes will be discussed further in this evaluation as the elements impacting the current sponge project design and implementations.

Many review papers have assessed the effectiveness (Xu et al. 2018; Leng et al. 2020) and evaluated the challenges and barriers to Sponge City implementation in China (Zheng et al. 2016; Jia et al. 2017; Li et al. 2017, 2020; Xia et al. 2017). A few studies have constructed their surveys based on analyzing the limitations and challenges of Sponge City implementation in China. We highlight them in this manuscript, in order to draw lessons and create a scheme policy model to avoid these shortcomings in the future. This review responds to the extent to which the Sponge City program has been successful in China, elucidating the gap between expectations and results by exploring the objectives, and performance of Sponge City. Cost-effectiveness evaluation is not covered in this review, although we direct readers to other papers in the same context to help readers gain a complete and global understanding of the subject.

Urban stormwater has long been a problem in major cities around the world, yet most solutions tend to use green technologies to manage this water. Traditional stormwater management seeks to efficiently collect and remove runoff from structures as quickly as possible. Therefore, they are designed to receive a limited amount of water that corresponds to design storms (Jones et al. 2005), causing downstream water quality deterioration, erosion, and flooding. During a heavy rain event, when the peak volume of stormwater exceeds the capacity of the sewage network system (combined or separated), conveyance systems overflow, and transfer the excess volume to nearby surface waters and water bodies. This, along with clogging, are the main cause of urban flooding problems, which resulted in the deterioration of water quality and jeopardized human health in municipalities around the world. Increased stormwater runoff leads to flooding, erosion, and degradation of ecosystem health. In addition, rapid urbanization worsens the problem, for example, US urban population increased from 65 to 82.7% between 1960 (Kim 2000) and 2020 (Statistica 2021), resulting in increased average annual urban runoff and flood events (Ahiablame & Shakya 2016). China and the US face similar water resource management challenges (drought, pollution, and water shortage). Both countries have developed similar water resource management policies and approaches to addressing these issues, beginning with the preservation of natural water resources and progressing to the construction of the largest water projects (such as the Grand Canal and the Three Gorges Dam in China, the Hoover Dam, and the California state water project in the US) and finally, shifting to more ecological solutions (He et al. 2020).

LID is referred to as ‘SuDS in the United Kingdom’ and ‘WSUD in Australia’. LID was pioneered in Prince George's County, Maryland, and Sponge City in China, and its basic principles are (Maliva 2019):

• Reduce impervious area.

• Disconnect impervious areas.

• Intercept stormwater before it comes in contact with impervious areas.

• Detain and infiltrate stormwater on-site, as close as possible to the source.

Results and benefits from LID practices have promptly shown as reduced in lower peak discharges, runoff coefficients, runoff volumes, and extended times to peak flow compared to traditionally drained zones (Damodaram et al. 2010; Hua et al. 2020). When China first adopted the Sponge City concept, the primary goal was to mitigate runoff, but this has quickly expanded to other goals such as urban runoff control, rainwater harvesting and reuse, water quality improvement, and ecological restoration (Jia et al. 2017). Controlling the first flush is critical to reduce the effect of road runoff on receiving water bodies, as it carries most rainfall pollutants (Liu et al. 2019c). The specific management measure for runoff pollution varied depending on the type of land use. Domestic Rainwater Harvesting (DRWH) system can mitigate peak runoff by 33% and volume runoff by 26%. Moreover, 40% of total annual rainfall can be stored in cisterns (ponds) and then treated in wastewater treatment plant for non-potable reuse.

The main objective of the LID concept is to achieve integrated urban stormwater runoff control, which is accomplished by replicating the natural hydrologic landscape and creating flow conditions that mimic the predevelopment flow regime, increased infiltration, and lengthening flow paths and runoff time (Damodaram et al. 2010; Ahiablame & Engel 2012; Yang et al. 2020), and its primary concern is to maintain that natural hydrologic cycle in urban areas during rainfall events. Hence, no model fits all cities, one of the most important goals of the LID technology is to resolve surface water quality and quantity problems caused by increasing runoff due to impervious areas.

This paper reviewed different journal papers and government publications by searching keywords including LID, Sponge City, stormwater management, and runoff mitigation in China, using mostly Google Scholar and Elsevier website. The selected articles were categorized based on their respective contents, simulation, review, or other. The papers were then treated, reviewed, and summarized separately to facilitate extracting data, and better present the information. Many previous reviews that assess the performance of the Sponge City (Eckart et al. 2017; Leng et al. 2020) focus mostly on stormwater pollutant removal rates and percentages of stormwater mitigation. Aside from flood mitigation, maintaining the same level of good performance over the project life cycle is also a critical component for a successful Sponge City project that cannot be overlooked.

This review explores the performance of Sponge City in detail involving its influence on urban life and human being from various aspects, with a focus on challenges of SPC as a concept, and on limitations that align with the optimal performance of SPC projects in China. LID facilities are not evaluated only based on their hydrologic performance but also based on their working process, this latter classifies LID into two categories: infiltration-based facilities such as permeable pavement and bioretention-based facilities such as rain gardens. Assessment of challenges and barriers during different implementation phases of SPC projects allowed for a more in-depth exploration of the existing simulation models. In the last part, the paper discusses the current policies to overcome the indicated problems, and proposes future policy orientations for the optimum Sponge City performance in China.

The SPC concept is subject to predetermined rules that have to be followed by all Sponge City projects, as the implementation of the concept is linked to many other factors such as rainfall and catchment characteristics. In addition, the criterion for assessing SPC projects are set by the ‘Assessment Standard for Sponge City Effect’, released in 2018 (Nie et al. 2020), which evaluates the SPC effect based on required and optional conditions.

Runoff control and flood mitigation are the two primary focuses of SPC concept. Hence, due to the uneven rainfall distribution, rainfall intensity and runoff volumes can vary according to the corresponding catchment. Furthermore, it is important to note that the volume capture ratio might vary among various SPC projects. Ideally, it should be set beyond the minimum threshold indicated in the zoning map for volume capture ratio of annual rainfall, which is specific to the project region. However, the planning and design of LID are based on catchment characteristics (infiltration ratio, groundwater table). Therefore, results from rainfall–runoff control studies that belong to the same limit on the zoning map are difficult to compare when assessing SPC effectiveness.

In this context, for the same rainfall event (rainfall intensity and duration) infiltration ratio differs from one facility to another. For example, as Table 1 shows, PP can effectively drain rainfall water even for large events, while RG or GR facilities are relatively less effective. LID stormwater facilities seem to be more effective at controlling the hydrological impacts of shorter return period events, while during large rainfall events, coupled green-grey-blue stormwater infrastructure system can ensure and maintain good performance (Leng et al. 2020).

Runoff reduction rates vary from LID facility to facility, however, they all promote permeability, runoff volume reduction, stormwater volume reduction, peak outflow reduction, and delayed time to peak flow. Eckart et al. (2017) have reviewed the performance of LID, and concluded that LID is more cost-effective than conventional stormwater management systems.

Professional engagement and public awareness can have a direct impact on the implementation and the effectiveness of the Sponge City initiative (Xie et al. 2020). Complying goals and objectives of SPC projects to local conditions and needs at a catchment scale, enhances the efficiency of SPC. Proper implementation of LID elements and regular maintenance are two important keys to optimum efficiency. However, optimal Sponge City design strikes a balance between cost and effectiveness. In humid areas, sponge cities tend to lay on bioretention facilities such as GRs and rain gardens, whereas in water storage and infiltration facilities, early rainwater treatment are the preferred measures in semi-arid and semi-humid areas (Griffiths et al. 2020; Shao et al. 2020). Due to its high infiltration rate and low cost, PP is utilized in approximately 30% of sponge areas. However, it may not be the optimal choice for regions with high groundwater tables.

Water quality improvement is one of the fundamental objectives of SPC construction. It is well known that urban runoff (including flooding) carries non-point-source pollution into receiving water bodies, causing damage to the environment. The official manual standard titled ‘Assessment Standard for Sponge City Effect’ published in 2018 by Nie et al. (2020), emphasized that SPC assessment should be based on at least one year of monitoring data, this latter should adhere to the project objectives set by local professionals engaged in the project. The effectiveness of the Sponge City concept in China can be evaluated on the basis of the efficiency of LID elements in reducing runoff and removing pollutants during their life cycle.

In Table 2, pollutant removal rates for different LID scenarios show results of pollutant removal rate studies based on LID scenarios. Obliviously, PP is widely adopted in numerous Chinese Sponge City initiatives, primarily due to its remarkable effectiveness and cost-efficiency. As an example, permeable pavement can mitigate stormwater runoff by more than 55% and remove most common runoff pollutants such as total suspended solids (TSS), total nitrogen (TN), total phosphorous (TP), and heavy metals, up to 89.6% (Huang et al. 2016). Huang et al. (2016) concluded in their paper that the removal of these pollutants is independent of climatic conditions, except for TN removal that tends to decline with a decrease in pavement temperature. In another study, by Hu et al. (2018) concluded that permeable concrete (PC) without clogging had the best performance in terms of flood mitigation and pollutant removal, and that, permeable interlocking concrete paver (PICP) is the least prone to being clogged. Furthermore, Liu et al. (2019b) in their paper have proven that even a GR facility, which is considered as an infiltration-based facility that promotes pollutant removal, can have a reverse impact on pollutant concentration if constructed with the wrong materials (substrate and plants).

Water quality improvement is an indicator that shows how well the LID meets its goal for all stormwater that reaches the LID site. In this context, Liu et al. (2016) evaluated and calculated the cost and benefits of different green stormwater devices (green space depression, porous brick pavement, storage pond, and their combination) by using the average service life method to evaluate the costs, the results have shown that the total average annual benefits of stormwater reduction and utilization by green infrastructures of the community are ranged from 63.24 to 250.15 thousand yuan, the benefit per cubic meter (m³) stormwater reduction and utilization is ranged from 5.78 to 11.14 yuan. While Rodríguez-Rojas et al. (2018) have compared and analyzed three types of PP, results have shown that their hydrologic performance can exceed that of vegetative SuDS reducing the operating costs of sewer systems and the flood risk. Several studies have examined the cost-effectiveness of LID practices, all have proven that, although the higher implementation cost, LID concept is long-term cost-effective (Liu et al. 2016; Mei et al. 2018; Qiu et al. 2020; Wang et al. 2020b; Wu et al. 2020).

The adoption of LID, in China, came to cover flood disaster consequences and their impact on social and economic life, control runoff, and storm pollution (Wu et al. 2020). Implementing Sponge City projects involves significant investment in the construction and furthermore in the operation and maintenance. However, SPC promotes cost-effectiveness in the long term, delivering cost-effective infrastructure that uses fewer natural resources than conventional drainage systems. Also, mitigating runoff at lower maintenance costs as less sediment goes into public drainage systems ensuring a long lifespan for it. Additionally, avoiding emissions, saving energy by reducing consumption of resources (Niu et al. 2010; Wang & Wang 2020; Xia et al. 2020), and most critically, reducing the effects of urban storms greatly lowers the cost of maintenance resulting from flood damage. Furthermore, the Sponge City concept treats stormwater as a usable resource rather than waste, preserving nature from emissions associated with the degradation of organic matter and treatments that require large energy inputs, usually carried out by the sanitation sector (Wu et al. 2020). Also, finally in addition to mitigating the urban heat island effect which reduces cooling costs in the summer and improves air quality (Wang & Wang 2020), SPC also creates attractive green spaces and water features that have a favorable contribution to the growth of tourism.

Despite this, it has been demonstrated that the economic value of sponge measures declined as the project's return period design rose (Wang et al. 2022a). However, project design and facility selection can have a considerable impact on the project's cost-effectiveness. That instance, constructing SPC projects based on the project's local conditions, such as soil types, groundwater table level, or climate conditions, would inevitably improve overall performance and raise economic advantages as from projects. In their case study, Oates et al. (2020) shown that Wuhan's Sponge City project is more than CNY 4 billion less expensive than a proposed alternative option to enhancing the city's resistance to flooding. Another case study in Xiamen, where Zhou et al. (2021) examined the economic benefits of investing in Sponge City infrastructure, revealed that, when all damages from flooding incidents are considered, SPC can outweigh the implementation costs.

LID technology provides a viable path for achieving a better environment, for both humans and nature. It contributes to the restoration of the natural hydrological cycle in urban areas by enhancing retention capacity, and as a result, moderating microclimate zones. Among the environmental implications of sponge cities or LIDs are (Liu et al. 2019c; Lund et al. 2019; Nowogoński 2020):

• Promotes aquifer recharge (infiltration supplies groundwater and maintains base flows to streams and wetland).

• Protects groundwater quality.

• Improves water and air quality.

• Reduces the pollution impacts of stormwater runoff.

• Urban heat island effect mitigation.

• Increased green space in urban environments.

Green spaces and plants help to create habitat, support urban animals, and boost overall ecological resilience. Furthermore, Sponge City strategies help to reduce the effects of urban heat islands, enhance air quality, and give possibilities for recreational activities and community engagement (Jaung et al. 2020). According to a recent study, addressing the shape and type of green space can reduce average land surface temperature by 1.32°C with only about 30% green space ratio (Liu et al. 2022). Moreover, the SPC concept prioritizes sustainable water management practices by promoting water conservation and resource efficiency. Rainwater harvesting systems, for example, collect and store rainwater for non-potable uses, reducing the demand for freshwater resources, and help alleviating water scarcity problems. Li et al. (2019) have estimated the individual benefits of each LID by quantifying the indicators weight of five LID design scenarios, and the results of the study have proven again the efficiency of SPC on the environmental, economic, and social aspects.

LID practices can change human behavior toward energy consumption, by reconciling people and nature, providing a harmonious and healthier lifestyle (Shen et al. 2020), such as protecting people and property from increased flood risk. In addition, creating attractive places where people want to live, work, and, play through integrating water and green spaces with the built environment. Improvements in human health resulting in a reduction in the frequency and severity of CSOs which results in better water and air quality and less pollution (Nguyen et al. 2020). Suppakittpaisarn et al. (2017) investigated the relationship between GI, GSI, and human health in his systematic review and observed that urban people who are exposed to more natural elements, such as LID devices (GRs, RGs, bio-retention swales), higher densities of trees around residential neighborhoods benefit from them noticeably (body, mind, and behavior aspects), as they are linked to cardiovascular health, healthier patterns of cortisol secretion, better pregnancy, and birth outcomes. According to a recent study on people's perceptions of SPC in China, 76.8% of the public is aware of the positive impact of SPC on social life and confirms its contribution to flooding management control (Zeng et al. 2023).

Sponge city initiatives frequently involve community engagement and participation in their planning and implementation (Wang & Palazzo 2021). Residents may be asked to participate in decision-making processes, provide feedback, and help design these projects. This increased participation fosters a sense of community ownership and pride, while also raising awareness about water conservation, sustainable practices, and environmental stewardship. Wang et al. (2022b) chose four cities in southeastern China to investigate residents' willingness to pay for SPC projects. The results show that most residents prefer green infrastructure to traditional drainage systems, and that they are willing to support the programs financially to some extent.

Urban heat island (UHI) occurs as a microclimate in densely urbanized areas that lack green infrastructure, as a result of albedo, cars heat emission, and air conditioning heat emission (Sun et al. 2020). These microclimates are typically hotter and more humid than the surrounding blue-green-grey areas; for example, the temperature difference between impermeable and permeable pavements can reach up to 15.8 °C (Liu et al. 2020). Permeable pavements are known to have a cooling effect during wet periods as they retain water for a few days after rainfall, due to their hydraulic conductivity properties. However, some types of PP perform better than others, such as reflective pavements, which absorb less solar radiation and emit less heat during critical times of the day, whereas permeable pavements only capture stormwater or excess irrigation water by allowing it to drain into the pavement and then evaporate (Li et al. 2013). Liu et al. (2020) have built, developed, and tested in the laboratory an innovative permeable pavement (IPP) called evaporation-enhancing PP. The results obtained in a bench scale investigation revealed that the IPP suppresses the surface temperature by as much as 9.4 °C, while other LID devices are mostly made up of a variety of plants that improve air quality and help reduce energy consumption.

The first phase in the implementation of an SPC project is data collection, and flood risk estimation which is the most challenging part. The estimation of rainfall threshold is fixed by National Assessment Standard (NAS), for each city according to the related needs, which can compromise the potential benefit of the project. When this fixed-rainfall threshold is overestimated, Yang et al. (2020) suggested considering non-stationary data for a proper threshold estimation, and higher potential benefit, also concluded that, preserving native vegetation and existing flow paths are basic keys for LID planning. While the design demands inspection and evaluation of the site hydrology, few studies consider groundwater direction and depth due to the difficulty in accessing such kind of data and the lack of or weakness of suitable simulation models. Additionally, the lack of public understanding and awareness of the Sponge City concept has a negative impact on its success, as evidenced by the lack of support and low public funding for Sponge City. Due to a lack of private sector involvement, adequate funding for operation and maintenance is difficult to maintain. Moreover, water projects are sometimes not financially viable because insufficient revenue is generated to cover the high initial investments as well as the operation and maintenance costs, which may cause future unsustainability of the field (Liang et al. 2020).

The lack of knowledge about sustainable stormwater management (SSM) leads to an underestimation of its importance and multiple benefits. Prioritizing quantifiable objectives within a short time frame, such as mitigating small storms runoff, prevents the implementation of new and unfamiliar measures (i.e., SSM). Implementing an SSM system that controls flooding requires a thorough understanding of the precipitation phenomenon, as better estimation of runoff leads to better performance of LID practices. As a result, the difficulty in obtaining data and the uneven spatiotemporal distribution of precipitation (Han & Wu 2019; Liu et al. 2019a; Randall et al. 2019; Wang et al. 2019; Chang et al. 2020; Hou et al. 2020; Wang et al. 2020b) influences the accuracy of rainfall estimation on a city scale. Consequently, small watersheds studies are more accurate, because they generate less uncertainty regarding the hydrological data of local conditions. Besides, it is more practical to substitute impermeable pavements with permeable alternatives in public areas like sidewalks, parks, and roads, thus emphasizing the widespread use of PP. However, in areas where the groundwater table is high or the soil is compacted, the infiltration rate can be drastically reduced (Starzec et al. 2020). As an example, Li et al. (2017) in their survey paper, have compared four SPC models, each one of them is located in a different climate region and have different soil and hydrologic characteristics, However, all LID models proposed for the four SPC were quite similar – with a rate of 38% to more than 70% of PP, which is justified by the fact that the design of SPC is limited by the lack of suitable vacant parcels to implement LID facilities, particularly in old areas, and urban centres (Oates et al. 2020; Qiao et al. 2020).

The guidelines policy in China promotes the implementation of LID in cities of newly developed districts and large cities. However, their implementation is more feasible in newly developed districts than in older urban sections due to the lack of vacant plots (Zhang et al. 2019). Yawen et al. (2020) suggested scientific decision-making and effective policy implementation, by involving technicians in decision-making. Or better yet, trained stockholders and decision-makers to introduce them to water management, allowing them to conceive more appropriate LID models that are tailored to each location for optimal performance (Li et al. 2017). All these factors enhance uncertainty in Sponge City design, implementation, and potential benefits, which is why people are aware of its success.

Furthermore, the implementation of LID elements is very expensive compared to the cost of constructing additional drainage supplies (Nguyen et al. 2020). As a solution, and in order to attract the private sector to participate in SPC financing and promotion, the government has set the public–private partnership (PPP) model which is based on involving private companies in building, operating projects under government control, to create a collaborative investment model between government and private companies. However, given the high cost of implementing the LID, private sector participation remains insignificant and insufficient. This goes back to the lack of a trustworthy assessment of Sponge City cost-effectiveness, which may dispel people's doubt about Sponge City success (Liang et al. 2020). Besides, to achieve the SDG ambition, new ideas about how sector funds from government, donors, and users are generated and allocated to create and sustain verifiable outcomes are required (Hope et al. 2020). Other policies enacted by the government to ensure funds and improve social acceptance, such as incentives on tax reductions, partially financing, construction permit, sustainability certification, are reviewed in detail (Li et al. 2017; Oates et al. 2020; Wang et al. 2020a).

Material provision and monitoring software of green infrastructure constitute another significant challenge toward a sustainable SPC construction, implementation, and operation as almost all LID materials except for permeable pavement are imported from abroad (Li et al. 2017).

Many challenges can occur post LID construction. These challenges are a result of systematic errors during the construction phase (such as inappropriate implementation of PP, or using wrong vegetation type for RG), or LID maintenance-related challenges such as PP waterlogging problem. This can be more challenging when LID elements are built on private land (e.g., GRs, green walls), as the owners may refuse or delay maintenance intervention. In this context, public acceptance and support can have an important effect on the life cycle of SPC.

Simulation studies are the most common topic in the area of Sponge City research works, as they can model all LID performance perspectives, and provide a visualization of the output of LID approaches for achieving the established objectives and goals (Ahiablame & Engel 2012). With these simulation models, planners can analyze different scenarios that help stakeholders in making strategic decisions. However, urban stormwater simulation bears many difficulties, such as the complexity of collecting precipitation data, defining existing urban drainage system impact, simulation models calibration, and so on.

Runoff estimation requires quantifying the hydrologic performance over time, from the beginning of the rainfall event to the end of surface runoff. Moreover, hydrologic performance is related to the climatic conditions of the study site, for example, temperature and solar radiation can affect the evaporation process, as well as atmospheric pressure and wind velocity. Estimating evapotranspiration and infiltration rates for each soil type and conditions are needed to separate runoff from rainfall. The separation between the runoff and the amount that infiltrates is called the production function.

Since LID's appearance in the US in the 70s, many models were developed to realistically simulate various urban water issues. Nowadays, with the adoption of LID principles under different concepts, more adapted simulation models have been improved to serve a better visualization of the targets. There are several methods for simulating surface runoff, including the curve number method, the Green-Ampt infiltration curve, the Philip infiltration curve, and the Horton infiltration curve methods. Therefore, the choice of the Sponge City simulation model depends on the availability of data and the objectives of the simulation. Despite the fact that most current models require a massive amount of data for their application, such as soil properties and compaction data, land use data, slope, antecedent moisture conditions, vegetation coverage, land management practices, long series precipitation data, and flow data, some methods, such as the curve number method, estimate runoff using only measured rainfall–runoff data. Table 4 summarizes some urban water simulation cases using simulation models commonly used in China. In order to obtain a more authentic runoff simulation, Lian et al. (2020) in their recent study used rainfall–runoff events data from 55 different sites across China to revise the runoff curve number to better fit with the Chinese natural environment. Accurate quantification of runoff is as important as calibration and validation of the simulation model of the study site. Nowadays, runoff simulation models are provided with more integrated methods and extensions that provide better quantification.

Because of high infiltration rates, infiltration-based facilities such as permeable pavements, GS, and porous parking lots can allow the intrusion of dissolved runoff pollutants into the soil causing contamination of groundwater (Qiu et al. 2020). Rodríguez-Rojas et al. (2018) confirmed this point by assessing infiltration rates for different types of permeable pavements, and have shown that the infiltration process starts very high, which makes PP a very good alternative for stormwater runoff control and flood hazard. However, groundwater can be contaminated with high rates of pollutants drained from the first flush. On one hand, higher infiltration rates are advantageous for stormwater management but not when water quality improvements are a primary objective for a particular location. On the other hand, PP efficiency decreases over time due to clogging, which may necessitate regular costly maintenance. In addition to the high construction costs compared to ordinary pavement, its use is limited in heavy traffic locations (mostly applied to parking areas and low traffic roads).

GS are also infiltration-based facilities, known for their high effectiveness in small to medium stormwater runoff volume reduction (Davis et al. 2012; Gong et al. 2019), as well as for their aesthetic and environmental benefits. However, they show poor efficiency in removing dissolved pollutants, and no efficiency when dealing with large storm events. In addition, GS are conceptualized to retain stormwater in order to decrease flow velocity, and then the stored water infiltrates, by high infiltration capacity, within 72 h of precipitation, to prevent causing bad odors and annoying insects. Besides, heavy stormwater may pose drowning hazards, when flow velocity surpasses infiltration capacity (Malaviya et al. 2019). In that case, the stormwater will press on both banks of the GS causing erosion. These latter are also limited by the topography of the site, as they require specific site conditions such as important slope and good soil capillarity.

Hence, in case of compacted soils or high groundwater tables, the infiltration capacity may be limited and fail to meet design expectations (Jin et al. 2022). For example, risks of pollutant intrusion to groundwater, or increased surface runoff generated by saturated soil is more likely to occur in areas with high groundwater table such as the eastern part of China (Yin et al. 2021). In this context, the design and the selection of LID for a specific SPC project are directly related to the site conditions. Therefore, the performance of a specific SPC project is equally compromised by spatial conditions and the applicability of a predicted cost-effective model (Schubert et al. 2017). Still, the use of infiltration-based facilities, such as PP and GR, with important land ratio, is not suitable if the main objectives of the project are focused on runoff pollutants control (Song 2022). Furthermore, in area suffering from water-scarce problems such as the southern part of China, the main goal is to use rainwater as a resource (Lancia et al. 2020). infiltration-based facilities are not the best option to use unless these are connected to drainage network, which engenders additional costs for the layers beneath the facility as well as for the connection to drainage network.

Bioretention-based facilities are a combination of plants, soil, filters, occasionally paired with grey structures such as a bypass or underlying pipes. Most components are sensitive and require extensive maintenance on a regular basis, resulting in higher operational costs (Köster 2019). The high maintenance demand is primarily due to waterlogging or sediment and organic matter accumulation. Clogged areas can obstruct water flow, reduce the effectiveness of pollutant removal, and require regular maintenance to function properly. Green stormwater control measures, in general, perform well in terms of pollutant control in small to medium stormwater but poorly in high intensity rainfall events (Si et al. 2022), while waterlogging risks are remediated by prioritizing to scenarios that show good control of peak runoff, and rainwater storage and regulation technologies. Furthermore, in areas with abundant water resources, for example, when runoff pollution control and runoff peak control are the primary goals, a combination of bioretention based facilities and infiltration-based facilities with better rainwater purification and peak reduction functions can be preferred (Liu et al. 2021a). Nonetheless, coupled stormwater management systems (green-grey) are the most effective scheme for preventing and controlling urban non-point source pollution (Qiu et al. 2020).

Another constraint of bioretention based facilities is the lack of available vacant parcels for implementation. It can be a significant challenge in densely developed urban environments. Due to limited available space, widespread adoption of bioretention as a stormwater management solution may be limited, which may affect SPC design (Wang et al. 2017). As a result, in existing dense urban areas, SPC implementation is very limited due to available land, justifying the excessive use of PP. Furthermore, LID practices such as bioretention areas or artificial wetlands necessitate adequate hydraulic capacity to manage stormwater runoff (Jin et al. 2022). Insufficient capacity may lead to overflows, reduced water retention, and decreased effectiveness of LID during high-intensity rainfall events (Jin et al. 2021). Additionally, for areas with high groundwater levels with rock layers, poor soil permeability, and steep terrain, necessary measures such as soil replacement, anti-seepage, and step setting should be taken prior to implementation to avoid secondary disasters that may increase construction costs.

The use of natural means to store, drain, and purify rainwater was a part of the house design in ancient China. Ancient homes were equipped with a central courtyard and a special terrace designed to allow the natural process for rainwater evacuation and drainage to the nearest water body. Nowadays, with the expansion of urbanization, we no longer find this type of architecture in cities, despite the fact that rainwater evacuation is a serious issue in urban areas (Yin et al. 2022). The Chinese government launched the SPC program to modernize and expand this tradition by pairing it with LID concept. The success of this policy process requires knowledge and contribution from various fields and organizations, as well as on-going monitoring. In this context, an assessment standard for Sponge City construction manual was released in 2018, by the Ministry of Housing and Urban-Rural Development (MOHURD), to set implementation and assessment rules on a national scale. Other means to ensure continuous control of the process are presented, that we will explore in more detail.

Sponge City is not a project with a limited life cycle; rather, it is a concept that one may be integrated into daily life or even part of a community's culture. Education is the finest way to introduce a new notion into society and urge folks to welcome, accept, and become aware of it, because public acceptance broadens the understanding of the Sponge City concept, policies, goals, and impact on people's daily life (Wang et al. 2022b), recognizing the need for adoption and support of these projects, particularly for the long-term objectives, which call for 80% of the municipal areas in China to be able to recycle 70% of incident rainfall by 2030 (Guiding Opinions of the General Office of the State Council on Promoting the Construction of Sponge City 2015) (Yin et al. 2022). Recently, the Chinese education system has begun to pay attention to integrating the concept of building sponge cities into the education system by organizing activities in schools and producing educational materials, such as ‘Design our own Sponge community’, ‘Sponge Castle Adventure (elementary version)’, and ‘Sponge City Exploration (middle school version)’ (Yin et al. 2022).

Since the launch of Sponge City program in 2014, a lot of research has been done based on case studies from different provinces in China, creating a database of different parameters and variables. Xu et al. (2022) have collected and created several charts and tables that help simplify the information and provide a clear image of the real situation of the Sponge City program, allowing research on the topic to move forward in the optimal direction with better accuracy and better results. Collecting site-specific performance data over several years, for example, reveals weaknesses and gaps in implementation, allowing design standards to be adjusted, technical manuals to be reformulated and above all implementation to be adapted to local conditions.

However, communication across project departments is critical during the technical design stage, as it is the initial stage of project success. Stakeholders and decision-makers must incorporate a deeper understanding of demands and target setting based on the hydrological characteristics of the city project in order to achieve optimal efficiency and cost-effectiveness. Furthermore, stakeholders and decision-makers should ensure constructive communication and collaboration between professionals involved in Sponge City project throughout all its life cycle, from the planning and design phase to the operational and maintenance process, in order to guarantee a smooth implementation and effective performance.

As mentioned earlier in the ‘Disadvantages and Limitations’ section, after several years of implementation, some pilot cities still experience flooding during extreme precipitation events, such as Wuhan Province in August 2020 or Henan Province in July 2021. As a matter of fact, sponge cities are not magical solutions that can eliminate flooding or deal with any precipitation event, but they can delay peak flow and lower its intensity, for a predetermined threshold at the planning stage. However, in case a flood occurs in a pilot Sponge City and the rain event that generates it is below the design threshold, then most likely, the design objective of the SPC was not reached due to systematic problems that can be solved methodically. These could be a result of errors in the design standards, for example, materials, construction, or other standards. In this regard, a study of current Sponge City design standards and their adoption conditions is essential for proper adaptation and error alleviation. Adopting wrong design standards may result in falsified simulation output and lead to an over- or under-estimation of needs.

The ‘Technical Guidelines for Building Sponge Cities in the People's Republic of China’ issued by MOHURD, in 2014, aim to incorporate the concepts of sponge cities into all urban planning, which appears to be more aspirational than reality. Integrating the Sponge City concept into urban planning entails considering the hydrological field and treating the Sponge City as part of the watershed rather than as part of the architectural planning. However, the SPC's current urban planning integrations are focused on transforming urban areas to be more permeable employing permeable pavement instead of non-permeable pavement and other LID facilities whenever possible. Hydrological correlation between neighboring watersheds, for example, or the distribution of sponge towns relative to watersheds, are rarely taken into consideration because they are usually built on public fields such as parks or sidewalks, which are typically chosen for architectural and economic aspects rather than hydrological or technical aspects. The policy of these standards guides, if correctly understood and executed, advocate a Sponge City design model at a watershed-scale where urbanization can harmonize nature. It is also worth to mention that the promotion of studies on the efficiency of LIDs built with local materials contributes to the development of a Chinese Sponge City model that only uses local resources, not only guarantee the benefits at an optimal cost but also create a new horizon for future urban water management.

The concept and implementation of LID facilities is new to the Chinese workforce. Besides, the green and grey combination of rainwater management systems generally requires a highly skilled and experienced workforce. Furthermore, building an RG or green roof, for example, is a complex process that requires close supervision to ensure that everything is carried out as planned, from soil placement and plants selection to main drainage underneath. However, mistakes committed due to the lack of experience might cause costly damages, affecting the efficiency and life cycle of the project. All these factors necessitate the need to establish a training program for the workforce involved in the implementation of SPC projects, not only to implement design plans, but local alternatives in terms of plants and other materials used should also be studied instead of importing species from abroad.

This review highlights the challenges, limitations, and barriers of current Sponge City projects in China based on performance assessment and provides a scheme model with various policies to address these shortcomings. Sponge City performance assessment can rely on two points: (1) Assessment of the implementation process, as Sponge City performance depends primarily on the appropriate implementation of the design. (2) On constructive communication between public administration and professionals involved in the project during all implementation processes. Moreover, the implementation of Sponge City in China is still facing many challenges, including the lack of data and uncertainty in simulation models. Successful implementation of Sponge City necessitates: (1) Thorough knowledge of the hydrology of the respective watersheds; (2) Accurate identification of the needs on a catchment scale; (3) Setting goals and objectives according to the needs.

Furthermore, the unavailability of a specific Chinese SPC design model is the main reason for its reliance on the western LID design model, which generates difficulties in providing materials at different stages, increasing implementation costs, and advantaging facilities that can be built with local materials (e.g., permeable pavement). Developing local models that address these concerns will open new research windows on resolving other issues such as hydrological connectivity between neighboring catchments that have been neglected in the implementation of most sponge cities.

The current implementation broadly prioritizes rapid urban runoff drainage at the lowest possible cost over other environmental implications of the SPC concept, such as penetration of stormwater pollutants into groundwater due to large use of permeable pavement. Also, design policy standards need to be updated, by adapting rain-flow simulation methods to the Chinese model, considering results from the short- and medium-term Sponge City implementation. Furthermore, taking into account the variability of rainfall distribution, soil characteristics, and the impact of climate change, for example, the cities in northern China, where climatic conditions and precipitation are quite different from those in the south, should adopt Sponge City concept based on their respective hydrological conditions. This review will help to better understand the current SPC implementation, in order to enhance future implementation and achieve long-term goals.

This work was funded by the National Natural Science Foundation of China (NSFC) (No. 52208084), the Hubei Provincial Natural Science Foundation (Grant No. 2021CFB005), the Startup funding of Wuhan University of Technology (40120684), and the Fundamental Research Funds for the Central universities (WUT:2023IVB062).

All authors contributed to the study conception and design. Material preparation, data collection, and analyses were performed by F.C. and X.Z. The first draft of the manuscript was written by F.C., and X.Z. has commented on previous versions of the manuscript. All authors read and approved the final manuscript.

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

The authors declare there is no conflict.