Bibliometric analysis of global performance and trends of research on combined sewer overflows (CSOs) from 1990 to 2022

Urban sewer systems can be classified into two main categories: combined sewer system (CSS) and separate sewer system. The CSS is a drainage system that collects domestic sewage, industrial wastewater, and urban runoff in the same channel and transports it to the wastewater treatment plants (WWTPs) (Li et al. 2022). Owing to historical circumstances, the majority of cities worldwide still maintain CSS in old urban areas. During periods of rainfall or snowmelt, CSS flow is positively correlated with rainfall intensity. When the flow rate in the pipeline exceeds the discharge capacity of the pipeline system or the treatment capacity of WWTPs, the mixed sewage surplus will be directly discharged into the receiving water body, which is referred to as combined sewer overflows (CSOs) (Brunsch et al. 2020). As urbanization intensifies, the current expansion of impermeable surfaces increases the amount of stormwater runoff, causing soil erosion and facilitating the transportation of a significant quantity of pollutants into the urban pipeline network (Baek et al. 2015). As urban nonpoint source pollution, CSOs are major contributors to the pollution of urban water environments; thus, the imperative to deal with and manage CSOs pollution has become increasingly pressing.

CSOs are a mixture of sewage and stormwater runoff, which have different properties. In addition, nonpoint source pollution caused by surface runoff scouring the street solid particles and sediments in the pipeline also converge into the CSOs. The diverse pollution sources of CSOs determine the complexity of the CSOs pollution (Peng et al. 2017). Controlling CSOs pollution necessitates a thorough understanding of the sources, destinations, discharge regulations, and influencing factors associated with CSOs contaminants. The main contaminants in CSOs include organic matter, nutrients such as nitrogen and phosphorus, suspended matter, heavy metals, pathogenic microorganisms, and other harmful substances (Gooré Bi et al. 2015). Furthermore, there is a growing concern about the presence of emerging contaminants (ECs) in CSOs. These ECs such as personal care and pharmaceutical products (PPCPs) (Poopipattana et al. 2021), microplastics (MPs)/nanoplastics (Di Nunno et al. 2021), antibiotics, and 0 (ARGs) (Eramo et al. 2017) are also entering the CSOs through the release of domestic sewage, urban subsurface scouring, and pipeline sediment scouring. CSOs are discontinuous, unstable, and highly polluted, thereby negatively affecting the functions of receiving water bodies, particularly low-flow water bodies, such as rivers (Chen et al. 2022). For example, the heavy metals contained in CSOs not only affect the growth and reproduction of aquatic organisms but also transfer and accumulate in aquatic organisms through the food chain, posing a potential risk to human health (Xu et al. 2018). Half a century ago, Germany, the USA, Japan, and other countries adopted various technologies to control CSOs through source control, pipeline system management, storage and regulation, and end-of-pipe treatment and developed various corresponding design specifications and standards. The increasingly prominent problem of CSOs and their ecological effects have become the focal points of current research in the field of environmental science.

Recently, researchers worldwide have conducted extensive research on pollution characteristics, pollution control technologies, ecological and environmental impacts, management technologies, and policies for CSOs (Table 1). Consequently, the number of research and review articles published related to CSOs has shown notable growth in recent years. These reviews provide an overview of the progress of monitoring (Vicory Jr & Tennant 1994; Weyand 1996), control technology (Weyand et al. 1993; Rönner-Holm et al. 2009; Tao et al. 2014), management technology (Peters & Zitomer 2021; Saddiqi et al. 2023), model prediction and evaluation (Roesner & Burgess 1992; Ma et al. 2023), disinfection (Wojtenko et al. 2001a, 2001b), and policies of CSOs (Wang et al. 2023). For example, Rizzo et al. (2020) reviewed constructed wetlands as a treatment method for CSOs and evaluated the performance of conventional water quality indicators, the removal of pathogens and ECs, and the effects of design and operating parameters on the overall removal performance of constructed wetlands. Petrie (2021) reviewed the impact of CSOs discharges on the concentrations of ECs in the environment and the treatment strategies for ECs in CSOs. Saddiqi et al. (2023) reviewed the application of technologies such as artificial intelligence, geographic information systems (GIS), and remote sensing in the management of CSOs. While these reviews summarized the research progress of various topics in CSOs, they do not offer a comprehensive analysis of the fundamental characteristics and challenges that encompass the research field of CSOs at a macroscopic level.

Bibliometrics is an academic methodology that relies on observable attributes of the studies. It employs statistical and mathematical methods to quantify these characteristics, including the quantitative relationships between the literature and the citations and co-citations, and depicts the current status and development trends of research in a specific field of study (Wang et al. 2021c). In contrast to conventional reviews that provide a comprehensive overview of the current state of research, bibliometrics focuses on prospective research directions and emerging trends in the research topic. Bibliometrics is now widely used in various research areas, such as bioretention facilities (Liu et al. 2023), sponge cities (Zha et al. 2021), stormwater management (Wu et al. 2019), and low-impact development (LID) facilities (Islam et al. 2021). The application of bibliometrics as an analysis method for CSOs studies has not yet been reported.

Considering the extensive body of literature pertaining to CSOs studies, the application of bibliometric methods to provide visual representations reflects the current status and development of the CSOs studies. In this study, based on bibliometric methods, we used the visual analysis software CiteSpace to review and visualize the current status of the research on CSOs through data mining, quantitative analysis, qualitative evaluation, time series analysis, and cluster statistical analysis. This paper presents a scientometric analysis of studies on CSOs and provides a useful reference for academics, researchers, and policymakers.

The existing Web of Science Core Collection (WOSCC) database (Thomson Reuters, USA) contains information regarding scientific content, impact, and collaboration for the most important and influential research literature and is recognized as one of the most comprehensive literature search platforms worldwide (Li et al. 2018). The search period for this study was from January 1, 1990, to December 31, 2022, and the search date was June 20, 2023, with ‘subject’ = ‘combined sewer overflow*’ or ‘combined system overflow*’ and ‘Document type’ = ‘article’. The WOSCC database contained 1,208 publications for the period between 1990 and 2022. Keywords with similar meanings were combined into a single keyword, such as ‘combined sewer overflow’ which stands for ‘combined system overflow,’ ‘combined sewer overflow,’ and ‘combined sewer overflows (CSOs)’.

Data from this study were processed and visualized using HistCite Pro, VOSviewer, and Origin 2018 software to summarize the global trends and development direction predictions for CSOs studies included in the WOSCC database for 33 years (from 1990 to 2022). The outputs included the number of publications, countries of publication, countries (regions), institutions, journals, author groups, and keywords.

HistCite Pro is a literature index analysis software that facilitates researchers in rapidly identifying the characteristics of literature (Garfield 2009). VOSviewer is a tool for building and visualizing a network of literature metrics based on citations, co-citations, or co-author relationships (Van Eck & Waltman 2010). The journal impact factor (IF) analysis for the estimation of the impacts of the journals was conducted using the ISI's Journal Citation Report (JCR) (Yang & Zhang 2013).

An analysis of journals publishing CSOs research can provide quick and accurate access to current authoritative journals on CSOs research. The JCR IF is a recognized means of characterization in academia, with a higher IF indicating a greater influence of the journal and a higher citation rate of its articles (Pérez-Hornero et al. 2013). Based on the results of the bibliometric analysis, CSOs studies published from 1990 to 2022 were distributed among 244 different journals, with Water Science and Technology having the most publications (269, 22.25%), followed by Science of the Total Environment (87, 7.20%), and Water Research (62, 5.13%). The top 10 journals, ranked by the number of publications, had a total of 617 publications, accounting for 51.03% of the total journals on CSOs (Table 4).

The journal Water Research had the highest total citation frequency, with values reaching 3,432, and a citation frequency per article of 55.35, ranking second among the top 10 journals. The number of articles published in Environmental Science & Technology was 23; however, the average number of citations per article was 69.17, and its IF₂₀₂₃ = 11.4, the second-highest journal in terms of the number of articles published, indicating that research published in this journal gained attention and readership from fellow scientists. Based on the 2023 JCR, six journals: Science of the Total Environment, Water Research, Journal of Hydrology, Journal of Environmental Management, Water Environment Research, and Environmental Science & Technology, were classified under the Q1 category, and the remaining four journals with IFs ranging from 2.7 to 5.8 were placed in the Q2 and Q3 categories in the citation report.

The cyan cluster mainly included the following keywords: ‘combined sewer overflow,’ ‘water quality,’ ‘nitrogen,’ ‘phosphorus,’ and ‘monitoring’. The yellow cluster mainly included ‘turbidity,’ ‘dissolved oxygen,’ and ‘organic matter,’ which represent the pollution characteristics of CSOs. The composition and distribution of contaminants in CSOs are always of primary concern. The main contaminants commonly detected in CSOs include suspended particulate matter, organic matter, heavy metals, nitrogen, and phosphorus (Yu et al. 2018). These contaminants are considered the primary control targets of the CSOs. As a primary contaminant in CSOs, particulate matter is an important carrier of contaminants, including chemical oxygen demand (COD), nutrients, organic matter, and heavy metals. It plays an important role in the water quality of CSOs and the chemical reaction of dissolved contaminants. Additionally, the presence of highly concentrated organic contaminants in CSOs necessitates a substantial amount of dissolved oxygen during the degradation process, which causes hypoxia in the water body when it exceeds the oxygen-enriched capacity of the water body and affects the normal survival of aquatic plants and animals, consequently affecting the aquatic ecology and resulting in the formation of black and odorous water bodies (Zhu et al. 2021). Oxidation-rich CSOs are rich in nitrogen and phosphorus, which promote the growth and reproduction of algal organisms, resulting in a brownish-green coloration of the water body and the formation of water bloom (Li et al. 2021).

The characteristics of different contaminants in CSOs significantly vary among countries owing to variations in precipitation characteristics, geographic conditions, intensity of human activities, and design parameters of drainage pipelines. With the increasing amount of fundamental data, it is essential to enhance the application of data models such as the environmental fluid dynamics code (Saharia et al. 2019), hydrodynamic models (Kouyi et al. 2011), and artificial intelligence models (El Ghazouli et al. 2022) to efficiently simulate water quality in CSOs and quantitatively evaluate ecological risks in receiving water ecosystems.

The blue cluster primarily included the following keywords: ‘Escherichia coli,’ ‘bacteria,’ ‘microbial source tracking,’ and ‘enteric viruses,’ which represent microbial contamination in CSOs. The primary source of microbial contamination in CSOs can be attributed to the deposition of precipitation on impermeable surfaces, such as roofs, and the presence of human feces, wildlife excrement (e.g., avian species), and domestic animal waste (particularly from canines). Of the total volume of CSOs, 5% contributes to over 99% of the pathogen load discharged into the receiving water. Pathogenic microorganisms constantly reproduce and spread using water as a medium, and can cause human intestinal diseases, respiratory diseases, myocarditis, meningitis, and other infectious diseases via fecal-oral and aerosol transmission. Pathogenic microorganisms in runoff and CSOs have become major sources of microbial contamination in surface water in urban areas (Sterk et al. 2016). Microbial source tracking (MST) is a highly efficient tool for identifying the source of pathogens in CSOs. Zeki et al. (2021) identified the main sources of microorganisms in estuaries through the detection of fecal-indicator bacteria and human-associated Bacteroides and found that rainfall had a moderately positive correlation with all MST indicators, indicating that CSOs are one of the main sources of microbial contamination in estuaries.

Furthermore, the presence of ARGs significantly contributes to increased pathogenicity resulting from microbial contamination. ARGs can be transferred between various microorganisms through horizontal gene transfer, thereby increasing the risk of water ecosystem contamination by pathogenic microorganisms (Cheng et al. 2020). Pathogenic microbial contamination has always been a challenging issue in surface water management, and CSOs lack sufficient attention as important sources of microbial contamination in surface water. The continuous development of molecular biology technologies, such as polymerase chain reaction, quantitative polymerase chain reaction, gene chips, and high-throughput detection technology based on second-generation sequencing technology, can facilitate better monitoring of the microbial contamination of CSOs and control them through corresponding strategies.

The purple cluster mainly included the following keywords: ‘wastewater,’ ‘contaminants,’ ‘micropollutants,’ ‘pharmaceuticals,’ ‘personal care products,’ ‘emerging contaminants,’ and ‘risk assessment,’ which represent the ECs in CSOs. Urban atmospheric and surface contaminants in the urban environment and on the ground surface are flushed, carried by runoff, and discharged into receiving water bodies through pipelines and CSOs. These ECs enter the environment and biomagnify through the food chain, posing larger risks to ecosystems and human health (Diepens & Koelmans 2018). In addition to conventional contaminants such as particulate matter, nitrogen, phosphorus, grease, and heavy metals, ECs include PPCPs (Hawkins et al. 2023), and MPs (Zhou et al. 2023). Unlike conventional contaminants, the hazardous characteristics of ECs are relatively hidden, and their impacts on human health and ecosystems are often chronic. Their long-term accumulation affects the growth, reproduction, and heredity of aquatic organisms and plants and seriously threatens human health through the enrichment of the food chain. Therefore, it is imperative to strengthen the detection and identification of ECs in CSOs and determine the appropriate EC control. In addition, there is a need to deepen the understanding of the migration and transformation mechanisms of ECs in pipeline systems and CSOs and elucidate the migration and transformation law of ECs in pipeline networks and their interaction mechanisms with the transformation of conventional contaminants (COD, nitrogen, and phosphorus). Further development of efficient adsorption and advanced oxidation technologies is essential for the removal and control of ECs from CSOs.

The red cluster mainly included the following keywords: ‘management,’ ‘model,’ ‘real-time control,’ ‘low impact development,’ and ‘green infrastructure,’ which represent the source reduction of CSOs pollution. Due to the harmful effects of CSOs on urban water environments, researchers have proposed three technological routes (source reduction, process control, and end-of-pipe treatment) for controlling CSOs pollution. These approaches involve source reduction, process control, and end-of-pipe treatment methods (Quaranta et al. 2022). During the rainy season, the main cause of CSOs is excess rainwater exceeding the interception capacity of the main pipeline; thus, it flows into the CSS. LID facilities, such as green roofs, concave herbaceous fields, and grassed swales can effectively facilitate the infiltration and absorption of rainwater, thereby reducing the amount of water entering the pipeline network. Additionally, filtration and interception methods, including the application of rain gardens, penetrating pavements, and bioretention facilities, can be applied to effectively reduce and remove the contaminants carried by runoff (Fu et al. 2019). In addition to the reconstruction of separate drainage systems, process control in CSOs management can be effectively achieved by increasing the interception ratio of the pipeline system and the number of storage ponds to transport the highly polluted stored mixed sewage through the pipeline network to the WWTPs for treatment.

The implementation of a real-time control system enables optimal utilization of the storage capacity in the pipe network, providing a more cost-efficient and efficient solution to solve the challenges of CSOs and waterlogging. The structure of the distributed optimal control system encompasses a real-time control mechanism for drainage systems, employing multiple agents that exchange information and make decisions based on local data. This distributed optimal control minimizes the data load on individual agents, effectively mitigating the high computational burden encountered in centralized control systems. Zhao et al. (2022) developed a distributed optimal control approach for regulating multipoint control units in urban sewer systems. For instance, in a sewer network located in Chaohu City, China, consisting of five pumping stations and 18 associated sewage or stormwater pumps, the total nitrogen (TN) loading released into the receiving waters could be maximally reduced by 37.6% through optimal control. Implementing suitable end-of-pipe measures by the contamination characteristics and hazards associated with CSOs is crucial for managing and mitigating CSOs pollution.

Furthermore, it is imperative to enhance national policy intervention, legal and regulatory mandates, and top-level planning and management strategies to enhance the management of CSOs and mitigate their adverse effects (Chen et al. 2019). With the development of sensor, measurement, communication, and computing technologies, data collection has become less challenging, paving the way for the establishment of intelligent monitoring networks and data infrastructures. By implementing technologies such as artificial intelligence, GIS, and remote sensing, it is possible to effectively manage runoff and CSOs, assess the degree of pollution and potential impacts of CSOs, and provide references and support for the construction of infrastructures adapted to climate change and urbanization (Saddiqi et al. 2023).

The green cluster mainly included the following keywords: ‘stormwater,’ ‘runoff,’ ‘sediment,’ and ‘pollution,’ which represent the study of pipeline sediments. Pipeline sediments are a significant source of CSOs contamination (Wang et al. 2021a). Studies have shown that pipeline sediments contribute between 30 and 80% of the pollution load in receiving water bodies during rainstorms. Specifically, the contribution of pipeline sediments to the TN content of the receiving water bodies ranges from 20.9 to 44.6%, whereas their contribution to the total phosphorus content ranges from 35.7 to 47.3% (Li et al. 2022). Pipeline sediments are also an important source of numerous ECs. Zhang et al. found that pipeline sediments are a major source of ARGs contamination in CSOs based on the ‘tidal’ flushing characteristics of CSOs contamination (Zhang et al. 2022). Urban pipelines are commonly silted with sediments, especially in CSS, which are easier to deposit during the dry season (Hua et al. 2022). The presence of sediments in pipelines can block pipelines, reduce drainage flow, induce pipeline corrosion, overload the drainage system, and cause premature overflow of the CSS. Additionally, sediments emit hazardous gases, such as hydrogen sulfide and methane. Therefore, the transportation of pipeline sediments during rainfall events is a significant contributor to environmental pollution (Chen et al. 2022).

Researchers have conducted in-depth studies on the structure and properties of pipeline sediments and numerous tests to explore the scour characteristics of pipeline sediments under different hydraulic conditions. Various data analyses have been conducted to predict the migration process of sediments under different conditions (Liu et al. 2021; Zhang et al. 2023). However, most recent studies have focused on the role of runoff and its initial scour on CSOs, whereas few studies have been conducted on the mode and degree of contribution of pipeline network sediments in CSOs. When the CSOs occur, a large amount of dissolved organic matter (DOM) will be transported into water bodies, which affects water eutrophication and ecosystem health. Therefore, determining the source contribution of DOM is of great importance to improve water pollution (Liao et al. 2021) provided a method to describe the mode and degree of contribution of pipeline network sediments to CSOs using the degradation potential index and the end member mixing model. Therefore, it is important to explore the contribution patterns of pipeline sediments to CSOs pollution to control it and the health of the urban water environment.

The orange cluster mainly included the following keywords: ‘CSO treatment,’ ‘disinfection,’ and ‘coagulation,’ which represent the control and treatment of CSOs. Since CSOs result from the mixture of dry season sewage, rainfall runoff, and sediment washed out from the pipeline, it has the characteristics of discontinuous, explosive and random pollution, and its short-term impact on the drainage system is stronger. Therefore, it is of great significance to take appropriate treatment measures to reduce CSOs pollution. The end-of-pipe treatment is the purification of pollutants at the end of the pipeline system to reduce the load of pollutants discharged into the receiving water, which is the most effective, most thorough, and fastest treatment to CSOs pollution.

The end-of-pipe treatment of CSOs primarily consists of ecological and technological measures. Ecological measures mainly refer to constructed wetland technology, which can reduce the speed of water flow, filter a variety of pollutants. However, due to the spatial constraint, it faces significant challenges in the engineering implementation. Technical measures include mechanical methods (e.g., cyclone separation, thin layer separation, sand filtration, and grille separation) and physicochemical methods (e.g., adsorption, coagulation, flocculation, and disinfection) (Masi et al. 2023). Cyclone separation is the most commonly used CSOs pollution control owing to its small footprint and low construction costs. Coagulation and sedimentation technologies are widely used to treat the pollution caused by CSOs. However, the dosages of chemicals in the coagulation and precipitation processes pose a challenge and can easily cause secondary pollution during rainstorms. To address this problem and enhance the pollution removal efficiency, researchers have proposed magnetic coagulation technology to control CSOs pollution. Furthermore, to reduce the microbial contamination of CSOs, typical disinfection methods are used, including chlorine, ozone, and ultraviolet disinfection (Crocetti et al. 2021).

Constructed wetlands are an end-of-pipe treatment that efficiently controls CSOs pollution. They have less construction investment, low operation cost, strong ability to handle pollution load. They can also be combined with the landscape in urban construction, which not only beautifies the urban environment but also regulate rainwater flooding and improve the ecological environment, which has a good prospect for development (Wang et al. 2021b). At present, research related to end-of-pipe treatment focuses on the simultaneous removal of suspended solids, nitrogen, phosphorus, organic pollutants, and microorganisms, while there is a lack of research on the removal of ECs. At the same time, the impact of treated effluent on the ecological restoration of receiving water bodies should be considered.

This study analyzed 1,209 articles retrieved from the WOSCC database between 1990 and 2022 in terms of the number of articles, countries/institutions, institutions, and journals based on bibliometrics and using the citation analysis software HistCite Pro and web visualization software VOSviewer. Keywords were used to determine research hotspots and developmental trends on CSOs. Based on the results of the analysis, the most important conclusions are as follows:

• (1)

  In terms of the number of published articles, there was an increasing global emphasis on CSOs research. The three most important research institutions were the Technical University of Denmark, the Tongji University, and the University of Sheffield. The USA has the highest number of publications related to CSOs, followed by Germany and the UK. Studies in this field were mainly published in journals such as Water Science and Technology, Science of the Total Environment, and Water Research. Using the VOSviewer software to analyze keywords, it was found that CSOs pollution sources, characteristics, risk assessment, and control measures are the current main research directions on CSOs. Pollution control measures for CSOs based on source control and ECs in CSOs are potential research topics.

• (2)

  The control of CSOs must be scientifically designed and planned. According to the technical guidelines and standards that can be implemented, the technical and economic analysis methods should be established to optimize the implementation plan and guide the control system design of CSOs. Formulate medium and long-term targets and plans for pollution control of CSOs in different cities, apply new concepts and methods of stormwater management, actively develop efficient control technologies and equipment, apply green infrastructure and gray infrastructure to reduce investment and improve control efficiency, and continuously monitor and evaluate the implementation effect. Optimize the operation and management, and finally solve the pollution problem of urban CSOs. The USA is one of the early pioneers in systematically implementing CSOs pollution control worldwide. The legal and policy system construction, basic research, and engineering practice of CSOs pollution control have been carried out for decades. Approximately 3.2 billion m³ of CSOs is discharged annually in the USA. Based on the 2000 estimate by the U. S. Environmental Protection Agency (EPA), at least $50.6 billion investment would be required to achieve 85% control of the CSOs, accounting for 27.9% of the total investment demand for drainage facilities in the USA. After decades of treatment, the pollution control of CSOs in the USA has achieved obvious results on the whole. For example, according to the statistics of 2014, 54% of the annual CSOs of seven states in the Great Lakes region have been controlled. However, CSOs pollution is still a problem that has not been completely solved in the USA, and the development level of different regions is also quite different. Many cities are still implementing more economical and efficient pollution control through various methods such as promoting green infrastructure (GI) and formulating long-term plans.

• (3)

  The design and construction of urban CSOs control systems require consideration of various factors, including real-time rainfall data, the configuration of pipeline networks, the population size, the scale of sewage treatment facilities, and the deposition of subsurface and pipeline networks. Furthermore, the composition of CSOs systems is highly intricate because it encompasses a range of technical measures, such as source control, pipeline network management, storage process control, and end-of-pipe control measures. Therefore, the accurate assessment of CSOs control systems is the basis for the correct understanding of CSOs in CSS and reasonable selection of the scale of CSOs pollution control measures. With the emergence of the concept of a ‘smart city,’ the application of computer models (e.g., artificial intelligence models) can provide substantial assistance in the planning, optimization, operation, and evaluation of CSOs control systems. The aforementioned strategy enables efficient planning, optimization, operation, and evaluation of CSOs control systems, thereby indicating a future direction for the design and construction of urban CSOs control systems.

Chen Shen contributed to the conceptualization and participated in writing the original draft and reviewed and edited the manuscript. Qingbang Yang contributed to the methodology and participated in writing the original draft; Zhonghong Li contributed to the methodology and software development, and visualized the preparation of the work.

This work was supported by the National Key R&D Program (No. 2016YFB0501505).

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

The authors declare there is no conflict.