Wetland water quality assessment of eco-engineered landscaping practices: a case study of constructed wetland parks in Hangzhou

Over the past decade, urban water systems have been affected by a range of detrimental impacts, including water quality and aquatic habitat degradation. Constructed wetlands (CWs) have proven to be effective in treating both domestic and industrial wastewater (Stefanakis 2019). Due to their impressive geo-ecological and landscape characteristics, constructed water quality treatment (WQT) wetlands have been extensively used worldwide for the treatment of wastewater. It has been observed that the pollutant removal rates of CWs are generally higher than other treatment measures when treating wastewater with high pollutant concentrations (80–99%) (Girts et al. 2012).

Constructed WQT wetlands are fundamentally different from artificial water features for they can function as a ‘filter’ and a ‘barrier’ (Knight & Kadlec 2000). The ‘water-soil-bacteria-plant’ system in WQT wetlands plays a complex role in diverse treatment processes, i.e., physical sedimentation, chemisorption, ion exchange, and biological uptake (Kurzbaum et al. 2012), whereas microorganisms, soils, substrates, and vegetation can interact synergistically with CW ecosystems to remove a broad range of pollutants (Greenway 2010). In addition to the characteristics of pollutants themselves, other local physical geographical factors can also affect WQT effectiveness, especially active storage retention time, wastewater treatment stages, size, layout, type, and density of aquatic vegetation (Acharya & Adak 2009). For practical ecological engineered landscaping (EEL) projects of constructed WQT wetland parks, several types of treatment stages have been typically applied, i.e., subsurface flow wetlands (SSF), vertical flow wetlands (VF), free surface flow wetlands (FSF), and floating wetland island (FWI). Primary treatment ponds (PTPs), aeration ponds (APs), and ornamental ponds (OPs) are also utilised in WQT wetland systems.

Researchers have also summarised different WQI evaluation criteria. Ammonia nitrogen, TP, TDS, and metals are the most common pollutants in domestic wastewater and industrial tailwater (Akcer & Dzemydiene 2020). It is believed that FWIs often perform TP and DO removal, while TN, TP, and TSS removal effects are also found (Oddsson et al. 2021). FWIs are also capable of removing targeted pollutants from stormwater. Seasonal WQIs, such as pH, salinity, total dissolved solids (TDS), electrical conductivity (EC), dissolved oxygen (DO), and temperature, have been investigated for adaptive EEL projects. In-situ WQI monitoring of urban river wetlands in the cold regions of northern China with their complete purification stages, i.e., horizontal subsurface flow wetland (HF), FSF, and FWI, while their effects on NH₃-N, TN, and TP removal have been indicated. Depth profiles of DO are found to be associated with several physical geographical factors, especially hydrological factors (Haddout et al. 2022). Effects of WQTs on plant-associated microbial communities are also primarily monitored on a laboratory scale (Clairmont & Slawson 2020). WQIs are also proven to be strongly related to habitat qualities and biodiversity conditions. The abundance of freshwater fish is often positively correlated with specific WQIs, i.e., TSS, water temperature, NH₃-N, DO, and TP (Koushlesh et al. 2022).

In recent decades, not enough attention has been paid to WQT-oriented EEL in high-modified urban CWs, and increasing anthropogenic stress has brought high negative impacts, i.e., water quality, biodiversity, and multi-functionalities. Though a few small-scale projects integrated with potential landscaping frameworks and measures with certain characteristics of EEL planning strategies have been proven to be highly sustainable (Olsson et al. 2004; Haron & Feisal 2020), most studies have omitted comparative analysis among constructed WQT wetland parks with EEL characteristics. Some geographical studies suggested that lakes ecological restoration projects may potentially be conducted by constructed integrated CWs to face severe environmental challenges, while extensive management and monitoring programmes are also necessary (Mostafa et al. 2022). Sustainable EEL designs with aesthetic characteristics have been implemented with proper development of nature-based solutions (NBS) and proven WQT technologies in some applicable geodesign-based EEL practices, especially ones being applied for river wetlands (Li et al. 2022), while typical WQT-oriented EEL in eco-parks design have seldom been studied by civil engineers, environmental geographers, and landscape architects (Zhao et al. 2015).

Hybrid wetlands comprising multiple WQT stages have seldom been applied in wetland EEL projects, despite their being economical and cost-effective (Deswal et al. 2023). Generally, there is an urgent need for assessments of EEL-oriented techniques for integrated WQT management in urban CWs. Correlation analysis among hydrological factors (e.g., temperature, water flow velocity, length of reach) and pollutant removal contributions of different WQT stages of hybrid WQT wetlands have rarely been conducted by researchers, which is significant for promoting practical design for WQT wetlands. Therefore, the research objective of this research is further in-situ research to identify the interrelationships between the hydrological factors and pollutant removal contributions of typical urban wetland parks at each WQT stage during different seasons, in order to reveal the WQT effectiveness of typical urban wetland parks with EEL measures in multiple stage WQT engineering projects.

The experiment was designed to identify the interrelationships among the WQT stages, WQIs (i.e., pH, DO, NH₃-N, COD_Cr, and TP), and the hydrological factors (i.e., temperature, velocity, and length of waterbodies) of four typical WQT wetland parks in Hangzhou City, P.R. China in three seasons (summer, autumn, and winter) at their WQT stages (SSF, VF, FSF, FWI, APs, and OPs). WQIs and hydraulic factors were monitored at each sample sites in-situ, respectively.

Water sampling was conducted on clear days during three seasons in 2021. Samples were taken under approximately 15 cm of water. Hydrologic indicators, i.e., pH, DO, and temperature, were measured in situ via the electrochemical probe by 86031 AZ IP67 instrument (manufactured by AZ Instrument Corp.) during the monitoring. Simultaneously, COD_Cr, NH₃-N, and TP levels were measured in situ by the LOHAND water testing field kit. These WQIs have been proven to be representative indices for WQT assessments and aquatic ecosystems in urban for human consumption. All the measurement procedures were performed in triplicate with mean ± standard deviation. Moreover, the water velocity was assessed by the time taken for each buoy to pass a fixed distance between each pair of adjacent points above the waterbodies.

Furthermore, several circumstantial factors were identified and recorded for additional WQT-oriented evaluations for sites, i.e., type of vegetation, geomorphology of the waterbody, length of reach, and landscape feature elements. These indicators show the correlation between each factor, i.e., temperature, water velocity, and length of reach.

The Pearson correlation test analysis was then processed through a Python statistical programme to quantitatively reveal the linear correlations between hydrological factors and pollutant removal contributions in different WQT stages, while the removal contribution of each WQI was quantitatively performed and explained.

At outlets of each treatment stage in Changqiao Creek Eco-Park, four sampling sites were selected. Descriptions of the per site are given in Table 2. Firstly, treated effluents from a buried treatment station were released into a PTP (site 1) and aerated via several influent cascades by an AP (site 2). An FSF wetland was then abundantly vegetated with aquatic, submerged, and floating plants (site 3). Finally, effluent flowed into integrated FSF + FWI wetlands (site 4).

At outlets of different water treatment stages in Shuimei Park, five sampling sites were selected. Descriptions of each site are listed in Table 3. Treated effluent from Linpin Domestic Sewage Treatment Plant is discharged into a PTP (site 1). Effluent flowed into a VF wetland at site 2. A narrow FSF wetland and influent cascades were set up to treat effluents (site 3). Eventually, FWIs were set up for further purification (site 4). Finally, effluents flowed into an OP for recreation (site 5).

Shuijing Park, which is also an attraction for visitors to enjoy the cooler microclimate and semi-natural waterscapes in summer, is well-known for its charming waterscape along the reach. At outlets of stages in Shuijing Park, five sampling sites are designated. Introductions of each site are shown in Table 4. An urban river flows into the reach at site one and flows into a small-scale VF (site 2), then a narrow FSF at site 3. Later, FWIs were also set at site 4, and sewage flowed into an OP at site 5.

Yangbei Lake Wetlands, located west of Hangzhou City, has been constructed by locals as fishponds since the Tang Dynasty. Since then, Yangbei Lake Wetlands have been expanded and have become a significant irrigation water source in Fuyang District, Hangzhou City until the 1970s. Since the 1980s, the water quality of the lake has been severely degraded due to industrialisation and urbanisation in surrounding areas. In 2019, an EEL-oriented wetland eco-park was constructed to ease pollution.

At outlets of various stages in Yangbei Lake Wetlands, seven sampling sites were selected. Water is charged into CWs at site 1 and treated at the sediment pond (site 2). A series of FSFs is set at sites 3 and 4. FWI is constructed at site 5. An AP with some influent cascades (site 6) and an OP (site 7) are also built. Detailed introductions per site are shown in Table 5.

In-situ measurements of hydrologic factors and WQIs have been measured quarterly at each of the four constructed WQT wetlands. WQIs were improved significantly after the continuous purification process of the WQT wetlands (Table 6). To evaluate WQT effects per treatment stage, differences in WQIs were also calculated for treatment stages. Removal rates of Changqiao in summer (COD_Cr for 43.07%, NH₃-N for 72.03%, and TP for 74.44%) are the most remarkable among all sites, which also exceeded the results of in-situ monitoring in homogenous small-scale sites (Czudar et al. 2011). From the statistics, we learned that there are undoubtedly significant differences in the WQT effects at diverse treatment stages in different seasons, while some remarkable phenomena and differentiators are discovered, which will be interpreted in Section 3.2.

A typical hydrogeochemical method, known as water quality indices (WQIs), has been adopted as a valid indicator of water environment assessment. According to The Standard Limits for Common Indicators of Surface Water Qualities in P. R. China (GB 3838-2002), WQIs of surface water are further evaluated as follows.

DO is one of the significant factors in preserving the balance of the water ecological environment, especially for maintaining the survival of aquatic organisms. All the WQT wetlands in the research have been indicated for their apparent WQT effects on increasing DO levels. DO levels at outlets of all sites satisfied the class II standard in both summer and autumn and the class IV standard in winter at least.

COD_Cr is a comprehensive index representing levels of organic substances in the water body. COD_Cr levels at outlets of four wetland parks satisfied the class IV standards in summer while almost meeting the class V standard in winter. Inter alia, Shuijing Park presented the most notable performance in COD_Cr removal contributions in summer (51.99%) and autumn (44.63%). In contrast, the least noticeable COD_Cr removal performances in both summer and autumn occurred in Yangbei Lake Wetlands.

Removal of NH₃-N in four eco-parks was detected. The main WQT stages in all parks have corresponding effects on NH₃-N reduction. The water quality at the outlets of three wetlands meets class I standards in summer. The most remarkable performance occurred in Yangbei Lake Wetland in summer (95.50%). However, there are also some distinctive profiles. For Changqiao Park, increases in NH₃-N concentration have occurred in summer for indefinite factors. Besides, mean TP concentrations at all sites ranged from 0.05 to 0.45 mg/L in summer, 0.02 to 0.45 mg/L in autumn, and 0.19 to 0.42 mg/L in winter. TP levels at the WQT wetland outlets almost met class IV standards in both summer and autumn but did not comply with class V standards at all outlets in winter. However, due to the limitation of in-situ survey, biological indicators have not been included.

For dynamic wetland ecosystems with WQT techniques, the EEL method operates in a ‘design-monitoring-evaluation-adjustment-management’ process, differing from the qualitative approach of other high-modified and artificial waterscapes. A Pearson correlation test analysis is necessary to determine the direct linear correlation between hydrological factors, including water temperature, flow velocity, and length of reach, and their contribution to the removal of various pollutants. The Pearson correlation test was conducted using a Python programme that utilised Matplotlib and Pandas packages. The analysis aimed to identify linear correlations between three key hydrological factors and the contributions to pollutant removal. The hypothesis testing method is adapted, and involves calculating the p-value of the non-correlation of two variables under the condition that the null hypothesis is valid, i.e., assuming that the two variables are not correlated. If p < 0.05, it indicates that there is a significant linear correlation between the two variables. The results are listed shown in Table 7.

We found some specific correlations between individual factors and the removal contribution of per WQT stages for all sites, which can be concluded as follows.

• (1)

  For stages of APs, length of reach is the major influential factor for COD_Cr (p = −0.868), TP (p = −0.826), and pH (p = 0.657). Velocities of water are related to increases in DO (p = 0.603).

• (2)

  For stages of FSFs, both water temperature (p = −0.539) and velocities of water (p = −0.522) are often related to TP removal contributions, while NH₃-N removals are affected by velocities (p = −0.749).

• (3)

  For stages of FWIs, increases in DO levels show positive relations to both temperature (p = 0.581) and velocities (p = 0.515), while COD_Cr removal contributions are related to temperature (p = −0.542). Removal of NH₃-N and TP are associated with the length of reaches (p = −0.688 and −0.633, respectively).

• (4)

  For stages of VFs, COD_Cr, NH₃-N, and TP removal contributions are affected by the waterbodies' temperature (p = −0.770, −0.789, and −0.571, respectively). Decreased pH is related to velocities (p = −0.861) and lengths of reaches (p = −0.812), indicating the valid capabilities to balance the pH of the water body.

• (5)

  For stages of OP, TP removal contributions positively relate to velocities of waterbodies (p = −0.772), while none of the other factors indicated more related aspects.

It is noticed that DO variations per treatment stages in wetland systems demonstrated their similarities. DO increasing trends are closely related to seasons (summer > autumn > winter). For summer, DO levels often rose during AP, VF, and FWS treatment stages. DO increases are processed during both stages of AP and FSF in autumn and winter.

The engineering parameters of wetlands, namely gradients, water velocity, and water temperature, appear to affect DO levels. Additionally, DO concentrations may be elevated by the photosynthesis of aquatic plants and the respiration of (micro)organisms. The reoxygenation reaction is an essential process for introducing oxygen into the water. Interactions between oxygen molecules in water frequently occur due to partial pressure gradients in the gas phase and a concentration gradient in the liquid phase. Therefore, more frequent turbulence in water bodies will decrease these gradients and improve oxygenation during the summer. However, stagnant water surfaces during autumn and winter will hinder oxygenation. Thus, extending the reach lengths of APs and FSFs will enhance reoxygenation in both seasons.

It is observed that variation trends of COD_Cr concentration at most treatment stages are almost similar, except for some stages of Yangbei Lake in winter. Removal trends of COD_Cr are related to seasons (summer > autumn > winter). The most significant removal effect occurred in FWS and FWI in the summer. The stage of AP in Changqiao presents positive results in removing COD_Cr (13.46%), while the AP in Yangbei Lake is explicitly non-effective (1.10%). However, distinctive profiles of COD_Cr removal contributions are found in autumn. VFs in Shuijing (6.78%) and Shuimei (14.33%) indicated more apparent effects. Removal contributions of FSFs and FWIs in all sites were respectively weaker. As for winter, removal rates of COD_Cr in all sites were minimised to the most limited ranges. Vegetation growth in the second half of all wetland parks was weaker than in the first half section. These phenomena can be elucidated as follows.

COD_Cr can be deposited, biodegradable, and transformed by assimilation through aerobic processes (Imfeld et al. 2009). During summer, photosynthetic rates were increased at all sites for higher temperatures and sufficient O₂ production. Biodegradation rates of COD_Cr reached a maximum with frequent microbial activities (Wang et al. 2023). Meanwhile, the O₂ transfer capacities of the depleted vegetation decreased in autumn. Plants and microbials were almost suppressed in winter of low temperatures. Integrated VF–FSF-resisted impact pollution loads of COD_Cr, while OPs presented negative, which may be caused by exogenous pollution as obstacles. Thus, COD_Cr removal efficiencies were inhibited.

NH₃-N removal primarily depends on plant uptake in constructed WQT wetland systems (Czerwionka et al. 2012). It is noticed that the total NH₃-N removal was seasonally dependent in all sites, i.e., summer > autumn > winter. Meanwhile, hybrid VF–FSF wetlands can enhance NH₃-N removal rates. Some anomalies in variations of NH₃-N levels were observed in Shuimei Park. NH₃-N levels in this park first decreased swiftly and then increased. FSF + FWI and OP had noticeable adverse effects on NH₃-N removal due to the scarcity of aquatic plants and the dominion of extraneous pollutants. It was also found that APs and some FWSs revealed inexplicit NH₃-N reduction.

Vegetations also optimise the hydraulic environment for microorganisms by releasing oxygen in root zones. Thus, dense vegetation attributed to noticeable NH₃-N removals in the summer. Total removal efficiencies are affected by the withering of plants and disappearance of microorganisms as decomposition of organic nitrogen can reduce NH₃-N removal rates. Furthermore, ammonia nitrogen is removed from substrates through filtration, adsorption, and sedimentation. On the surface, microorganisms may also conduct nitrification and denitrification processes (Jiang & Chui 2023). Additionally, isolate-set FWIs tend to be ineffective at promoting nitrification due to the constraining effects on oxygen transportation. Root zones can create microenvironments that support microbial communities and provide organic carbon sources for aerobic bacteria, leading to an increase in the structure and abundance of the microbial community (Di Luca et al. 2019). Studies have suggested that integrating FWIs with VFs and/or FSFs can be an effective method of improving WQT efficiencies.

Phosphorus is one of the reliable chemical nutrients for algae to grow in water. Water-soluble reactive phosphorus (SRP) is absorbed and transformed by plants. TP was removed through three parallel pathways, i.e., substrate fixation, plant uptake, and microbial action. Each component contributes differently, i.e., substrate > vegetation > microbes (Wang et al. 2017). Total NH₃-N removals are also observed to be closely related to seasons. It is indicated that TP removal rates of VF, FSF, and FWI in winter are much lower than those in summer. These phenomena are further explained as follows.

Absorption and sedimentation are both dominant for TP removal. Rahman et al. (2020) have noted that TP removal depends on multiple influences, i.e., sorption and complexation by vegetation, sedimentation, and biochemical action by microorganisms, while the deposition of suspended particles removes particulate P. Vegetations enhanced treatment efficiencies of WQT wetlands by absorbing and releasing O₂ and transforming organic matters to promote microbial degradation. O₂ release from roots and secretion of organic matter often promote microbial degradation. It has been suggested that the accumulation of organic P and the concentration of TP in aquatic plants aligns with prior research (Wang et al. 2023). The growth and reproduction rates of phosphorus-accumulating bacteria are also affected by temperature. In semi-anoxic environments, DO concentrations in wetland beds decrease and thus limit bacterial growth. As a result, there is a shift from phosphorus uptake during winter to phosphorus releases.

Several rules can be concluded for the correlation between hydrological factors and WQIs.

• (1)

  Water temperatures ultimately impact WQT effects for all stages.

• (2)

  Lengths of reaches are frequently related to the removal processes of TP. Longer reaches with meanders will be of help to promote TP removal.

• (3)

  NH₃-N and decreases in pH. Velocities of waterbodies are frequently related to DO, TP and decreases in pH, which occurs with interactional effects of VF > FSF, and AP > FSF.

• (4)

  Several EEL measures effectively promote COD_Cr, NH₃-N, TP removal, and DO level rises, i.e., increasing lengths of AP, VF, FSF, and accelerating AP velocities, FSF, FWI, and OP.

• (5)

  The effluent concentration decreases from the previous stage to the latter, irrespective of the season. Removal of NH₃-N occurs primarily in the first half section, mainly in the first 1/4 section in summer, and gradually decreases among later stages in winter. TP occurs primarily in the first 1/4 section in summer and the first half section in winter. The plant role is high in the first section. Eventually, the phosphorus is removed by plant harvesting and substrate replacements.

• (6)

  Nutrient abundance in the water is explicitly higher at the inflow, thus, the rate of microbial degradation of organic matter is higher, resulting in a noticeably faster COD_Cr reduction. Once the sewage flows into the latter part of the wetland, the organic matter degradation rate reaches a certain level and the reaction rate decreases.

• (7)

  During preliminary aeration, numerous microorganisms proliferate in the form of activated sludge. If tailwater is directly discharged into the VF or SSF, it may obstruct the flow paths. Therefore, the AP or FSF should be installed between the AP and SSF/VF to avoid blockage.

Most of the treatment stages investigated in the study were discovered to have contributed to the improvement of WQIs, especially as their successful outcomes were recorded during the summer season, which aligns with findings from another in-situ monitoring cases (Li et al. 2009; Wang et al. 2021a, 2021b). However, significant differences in winter were observed among the wetland parks despite the noted hydrological factors. Compared with other sites, Shuijing Park and Yangbei Lake Wetlands have sustained their removal rates, which were also attributed to their different EEL measures. These perceptible differences can be explained as follows.

The activities of nitrifying and denitrifying bacteria have an impact on the removal rate of NH₃-N. The denitrification rate varies seasonally due to physiological and biochemical changes that occur in the microenvironment surrounding the root zones of vegetation. A lower rate of denitrification was observed during winter, which is consistent with other assessment programmes in southern China (Wang et al. 2021a, 2021b; Jiang & Chui 2023). Therefore, while the nitrification efficiency may remain high in low-temperature conditions, the denitrification rate limits the overall nitrogen removal in the WQT wetland, and vice versa. Both parks have larger planting areas. As a result, the vegetation roots have a greater ability to transport oxygen, enabling the maintenance of aerobic environments throughout the winter. The substantial surface area of the Yangbei Lake Wetland and the Shuijing Park, combined with their enhanced fluidity, creates an insulating effect that provides the perfect environment for the growth of aquatic plants during winter. Additionally, stages of AP exhibit more favourable eutrophic conditions, while ample DO in the upper strata of the water in VFs creates a suitable atmosphere for nitrification.

It is observed that the removal of COD_Cr and TN in FSFs is positively correlated with the inflow concentration, while the removal of TP is negatively correlated with the sewage inlet concentration, which is consistent with earlier studies (Mohd Noor et al. 2014; Zhu et al. 2023). Nonetheless, the effluent discharged from the FSF displays a significant concentration of nitrate nitrogen. The oxygen in the SSF predominantly originates from the oxygen released by the roots, and reduced oxygen levels may impede the efficiency of the nitrification process. Deeper sections within the SSF bed profile differ and alter from an aerobic to anaerobic environment, thereby promoting denitrification, as reported by Khan et al. (2022). The oxygen in the SSF predominantly originates from the oxygen released by the roots, and reduced oxygen levels may impede the efficiency of the nitrification process. However, the system's sustainability is upheld through VN-denitrification during times of insufficient carbon source. Nonetheless, it is important to consider that the VF beds are deeper and more prone to clogging during maintenance.

The findings demonstrate that the WQT contributions of all parks during specific seasons and stages were poor. Although VFs successfully completed nitrification, they were not reliable for denitrification reactions. Therefore, hydrologic engineering measures should be taken to optimise external physical geographical conditions of wetlands. To achieve sustainable and resilient landscaping, EEL projects for wetland parks must consider the various aspects of geo-ecological principles. Landscape architects should design suitable treatment facilities with proper composition of landscape elements, such as substrates, flow patterns, and vegetation.

By adopting the WQT measures, most treatment processes can be integrated into hybrid treatment stages of wetlands. Firstly, current deflecting elements, e.g., swales and microtopography, can promote the differentiation of flows, enhancing aeration and promoting higher DO levels. Secondly, by constructing low ditches along shores, lower sections of the streambed can be widened to promote the natural migration of (micro)organisms, which will enhance the removal of NH₃-N and TP. Thirdly, it is believed that VFs are usually more effective in COD_Cr removal and nitrification than HFs and FSFs in existing monitoring programmes (Abou-Elela et al. 2013; You et al. 2023), while VFs in this study have failed to present their ideal effects, probably for the limited tidal VFs prohibiting sufficient oxygen from penetrating, for their limited areas and inadequate detention time, which should be optimised by proper local-scale geospatial arrangements. Besides, waste-derived substrates are viable alternatives having fertilising effects with the potential for nutrient recovery, which is suitable for growing ornamental plants in wetland EEL projects (Sharma 2023).

The conspicuous removal effects of organic matter, i.e., N and P, is one of the distinguishing features of WQT wetlands. It is noticed that higher planting with higher densities is effective in reducing TSS and COD_Cr and increasing pH in the aquatic environment, which is essential for improving water transparency, which has correspondent with existing studies in the southern region in P.R. China (Xiang et al. 2022). Different plants should be arranged accordingly for distinctive WQT stages. Some native plants in the southern region of P.R. China often present effects on dephosphorization, i.e., Hydrocharis dubia in winter and Eichhornia crassipes in summer (Wang et al. 2021a, 2021b). It is found that adequate aquatic plant communities are effective in VFs. They are also effective for SSF wetlands. Additionally, higher planting density led to higher WQT effects, corresponding to studies conducted in a horizontal SSF (Shruthi & Shivashankara 2022; Rohaningsih et al. 2023). Emergent plant communities provide more shaded areas. The presence of floating vegetation in the FSFs and FWIs will enhance biodiversity while preventing the growth of algae. According to the results of WQI assessment, we recommend Phragmites australis as the most proper aquatic plant for total nitrogen removal from sewage in all the parks. The roots of this plant absorb pollutants in water, transport oxygen, and create ‘anaerobic-hypoxic-aerobic’ microenvironments.

Besides, lower water temperatures primarily affect the removal rates of VFs. Therefore, we recommend improving the insulation of the WQT wetland by covering the surfaces of the SSF and VF with insulating material during the winter period. Consistency in these measures will help improve the performance of the wetland. Additionally, to promote stable WQT effects, particularly for smaller-scale sites, we advise increasing the depth of the primary treatment unit and extending the hydraulic retention time.

Located in Hangzhou, four eco-parks in the study are adjacent to the core water systems of the city, containing many cultural and natural attractions. Non-point source pollution loads can be recurrently reduced by utilising the filtering and wastewater interception functions of constructed hydro-ecosystems, which can improve riparian landscapes (Huang et al. 2022). In a local geo-ecological scale, landscape architects should concentrate on integrating diverse functional layers, i.e., substrate layer, vegetation layer, waterbody layer, and WQT stages.

For adaptive ecosystem service management, EEL-oriented goals for multifunctional WQT wetlands should not be assessed by their state of completion but rather for the sustainability and resilience to deal with future uncertainties. Although the water quality improvement effect of the four eco-parks was significant, the landscape effect at some nodes is recommended to be optimised. Strategies of multifunctional landscape design should be optimised, which will help to enhance divergent functions to limited spaces with EEL measures for their management. We believe that proper-designed constructed WQT wetland parks should present multiple characteristics, i.e., environmental, sociocultural, and economic benefits. The four benefits mentioned above are summarised in Table 8. It is indicated that Changqiao Park and Shuimei Lake Wetland offered more generous social, cultural, and economic benefits. The EEL-oriented design of Shuijing Park and Yangbei Lake Wetlands should be improved to seize further sustainable development opportunities, i.e., natural education and social interaction.

From an ecological perspective, the author suggests that the findings will be beneficial for improving water quality in WQT stages as one of the ecosystem services provided by EEL projects of WQT wetland parks. Landscape architects and civil engineers are recommended to adopt some principles to optimise WQT wetlands. The diversity and landscape character of plants should be considered in practical EEL-oriented wetland design. Emergent, floating, and submerged plants should be arranged in communities based on their effectiveness in removing contaminants. Hybrid wetlands should also be arranged based on their WQT efficacy. During WQT wetland EEL projects, adaptive management of dynamic EEL should be considered. By enhancing EEL variability, semi-natural wetland ecosystems can progress from an initial state to a more stable ecosystem and potentially reach a stable state. Furthermore, WQT wetlands have been identified as highly promising landscape tools as they can serve multiple biodiversity-friendly functions and create semi-natural habitats within urban wetlands.

Some limitations of the study are also noticed, listed as follows.

Firstly, in terms of the scope, the research focus on the practical EEL projects of urban WQT wetland, whereas EEL projects for other types of urban waterbodies, including rivers, streams, and reservoirs, are not considered. Also, for EEL projects in other regions of the world, specific empirical studies should be further conducted by multi-discipline researchers.

Secondly, the indicators in the empirical experimental methods did not include biological indicators, especially for plant diversity. Plants in the CWs often offer their functions for water quality maintenance, and the plant diversity in the WQT stages has a direct relationship with water qualities and hydrological conditions (Júnior et al. 2023). Therefore, further research on the interrelations among the plant biological indicators and the WQIs is needed.

Thirdly, other factors such as water depth and carbon source (Zhu et al. 2023) may influence nutrient concentrations in CWs, which should be considered for future homogeneous in-situ assessments. Despite WQI monitoring of different types of specific CWs (Salah-Tazdaït & Tazdaït 2023), further research into the hybrid wetland system's mechanism is required.

Finally, for large-scale urban CWs and river wetlands, the dataset of WQIs and hydraulic factors could further integrated into a geographic information system (GIS). This integration would also be applicable for landscape architects and environmental geographers to analyse and visualise the data using geodesign techniques and geoinformatics methods (Li et al. 2022).

To identify the WQT effects of constructed WQT wetland parks for practical EEL projects, WQT performances of various treatment stages of four typical WQT wetland parks in Hangzhou City have been studied, all of which are indicated to achieve integrated EEL performances. In-situ measurements affirmed that all WQT wetland parks contributed to the removals of COD_Cr, NH₃-N, and TP and rises of DO. The increasing DO, and decreasing COD_Cr, NH₃-N, TP, and pH levels are explicitly related to seasons (summer > autumn > winter). Hydraulic factors, including the velocity, the temperature, and the length of waterbody, have affected comprehensive treatment performances in all seasons. Guidance for urban constructed wetland EEL projects should be further optimised, i.e., optimising the EEL measures according to WQT principles, adapting appropriate planting designs, and integrating with multifunctional design. These findings will be helpful for improving the efficacy of WQT wetland parks as one of the ecosystem services provided by the urban sustainable EEL projects.

This research was supported by the National Natural Science Foundation of China (Project #51208467); the Scientific Research Foundation for the High-level Talent Introduction, and the Zhejiang University of Technology (Project #118001929).

All relevant data are included in the paper.

The authors declare there is no conflict.