The importance of mass balance in life cycle assessment of nature-based solutions for wastewater treatment: key learning points from a case study

Nature-based solutions (NBS) are approaches that use natural processes and ecosystems to address various environmental, social, and economic challenges (European Commission 2015). These solutions aim to protect, sustainably manage, and restore natural or modified ecosystems to provide benefits such as enhancing biodiversity, improving water quality, reducing flood risks, mitigating climate change, and promoting human well-being. NBS include a wide range of strategies, such as green infrastructure, ecosystem restoration, and the sustainable use of natural resources, tailored to meet specific local needs and conditions (Nesshöver et al. 2017). For the last several decades, various new NBS technologies have been developed across all continents (Castellar et al. 2021). In particular, NBS are being promoted as a way of improving the resilience and sustainability of cities in the face of a wide range of environmental challenges (Ershad Sarabi et al. 2019). In the field of urban water management, they are seen as an alternative to more centralized and ‘grey’ treatment in order to move towards more sustainable management of water and biogeochemical cycles (Masi et al. 2018; Flores et al. 2019).

In NBS for water management, certain installations prevent flooding by diverting stormwater flow or enabling faster infiltration into the soil. This is particularly the case for infiltration basins, retention ponds, swales, and rain gardens (Woods-Ballard et al. 2015; URBANGREENUP 2018; UNALAB 2019). Treatment wetlands can also improve urban water management and treat wastewater before its discharge into the natural environment by replicating natural processes occurring in natural wetlands involving vegetation, soils, and the associated microbial assemblages (Dotro et al. 2017; Stefanakis 2019).

To accurately assess the environmental impacts of NBS technologies, it is crucial to consider both local impacts during the ‘use phase’ (e.g., water quality improvement) and impacts from other life cycle phases (e.g., construction and end-of-life management) (Corominas et al. 2020; Larrey-Lassalle et al. 2022; Babí Almenar et al. 2023). Life cycle assessment (LCA) is a standardized and widely used method for evaluating the environmental impacts of products and services from a life cycle perspective (Hellweg & Milà i Canals 2014). It is a reference method for the environmental assessment of technologies in the field of water management (Corominas et al. 2020; Parra-Saldivar et al. 2020; Larrey-Lassalle et al. 2022). Numerous water-related NBS technologies have been analysed using LCA, including swales and ponds (Byrne et al. 2017; Bixler et al. 2019), rain gardens for rainwater or stormwater management (Flynn & Traver 2013; Petit-Boix et al. 2015; Vineyard et al. 2015), and treatment wetlands for wastewater treatment (WWT) (Roux Boutin & Risch 2010; Fuchs et al. 2011; Pan Zhu & Ye 2011; Flores et al. 2019; Resende et al. 2019; Risch Boutin & Roux 2021).

When focusing on the studies for WWT, system boundaries always encompass the life cycle stages related to construction, maintenance operations, and end-of-life of materials, as well as emissions to water in the water exiting the system. Therefore, input and output water qualities are assessed. Greenhouse gas (GHG) emissions (CO₂, CH₄, N₂O) during the use phase are sometimes included, based on measurements or the International Panel on Climate Change (IPCC) emission factors (Fuchs Mihelcic & Gierke 2011; Pan Zhu & Ye 2011; Resende Nolasco & Pacca 2019). Some rare studies also include sludge spreading, with its fertilizing effect, and the related soil and air emissions (Roux Boutin & Risch 2010; Flores et al. 2019; Risch Boutin & Roux 2021). However, these studies highlight a lack of harmonization in the definition of system boundaries and of robustness in the calculation of mass balances. Most recent litterature review papers (Corominas et al. 2020; Parra-Saldivar et al. 2020; Larrey-Lassalle et al. 2022) proposed a set of guidelines for carrying out the LCA of WWT technologies. It is thus essential to carry out a complete mass balance to estimate the environmental impacts of a technology for all the environmental compartments, i.e. air, water, and soil (Corominas et al. 2020; Larrey-Lassalle et al. 2022) as it is widely acknowledged in grey technologies for WWT (Heimersson et al. 2016).

This paper aims to implement a complete LCA of an NBS water treatment technology, from the construction phase to end-of-life and sludge management. Particular attention will be paid to calculating a balanced mass flow for carbon, nitrogen, and phosphorus in the three environmental compartments. A French vertical flow (VF) wetland (treatment wetland) for WWT is used as proof of concept. Comparing data and methodological choices regarding system boundaries will enable the formulation of recommendations for evaluating NBS versus grey technologies in WWT.

The environmental assessment approach is based on the following four standardized phases of the LCA framework (ISO 2006): (i) goal and scope definition, (ii) life cycle inventory (LCI), (iii) life cycle impact assessment (LCIA), and (iv) interpretation of results with a focus on the challenges of taking into account a complete mass balance of C, N, and P flows in LCA of NBS for WWT.

This study aims to carry out a complete LCA of an NBS for WWT and discusses methodological choices related to system boundaries and data collection related to C, N, and P mass balance. A comparison with a grey technology is also proposed for interpretation and discussion.

The functional unit (FU) was ‘to treat 1 m³ of raw wastewater in France, with an abatement rate of 80% on the COD’. This corresponds to an FU conventionally used for WWT, integrating the regulatory requirements of small treatment plants in France (Directive Eau Résiduaire Urbaine 1991). Therefore, the rationale for this choice is to ensure comparisons with conventional solutions, i.e. grey technologies, such as activated sludge.

According to the recommendations of Corominas et al. (2020) and INRAE Transfert (2022), a wide perspective is adopted for system boundaries that encompass the WWT technology, and sludge treatment, transport, and application (Figure 1, System B). Since sludge is regarded as a coproduct of WWT, ISO 14044 recommends that system expansion be applied to estimate the impacts of this coproduction, including the integration of avoided mineral fertilizer production (Figure 1, System A). For WWT technology, construction, including excavation, materials for the filters and surrounding equipment, and operation and maintenance of the filters during the entire lifetime, up to the dismantlement of the NBS are considered. A particular focus is placed on the use phase through N, P, and C flow mass balances and sludge spreading.

Figure 1 illustrates the sequential treatment phases of the French VF wetland (two stages, with three and two compartments, respectively), while Table 1 summarizes the key structural properties of the construction for the first and second stages. The presence of filters in parallel for each treatment step allows rotation in feeding, leaving some filters resting. Indeed, wastewater is poured during a feeding period (3.5 days) in a first filter. Then, a technician changes the flow direction to feed another filter. The resting period is essential not only to complete the treatment of the interstitial water remaining in the filter, but also to partially degrade and mineralize the surface deposit formed during the feeding period, thus partially restoring the infiltration capacity. The second stage has only two compartments, which are successfully fed for 7 days and then left to rest for 7 days. Both are covered with a geomembrane and a geotextile to ensure waterproofing.

The lifespan of such a constructed wetland is 30 years. It is assumed that the previous land occupation was pasture or meadow. After the dismantlement, the land occupation would come back to this initial state. The volume of excavated soil, computed from the filters geometry, was used to estimate machine operation and energy consumption for construction and dismantlement, assuming an excavation rate of 110 m³ h⁻¹ (ACV4E Software 2018; Risch Boutin & Roux 2021). Each hour of construction or dismantling makes use of a dump truck for 0.7 h, a mechanical shovel for 0.7 h, a dumper for 0.3 h, and an excavator for 0.3 h (ACV4E Software 2018; Risch Boutin & Roux 2021).

Surrounding the French VF wetland, a grass area of 3,000 m² has a buffer role and holds an operating room, a storm overflow, inlet and outlet channels, and some storage tanks. These structures are made of 25 t of reinforced concrete and 22 t of lean concrete, assuming that there are local materials for the transportation.

The various stage feed structures (self-priming siphons), as well as the manual bar screen, have been taken into account through the consumption of raw materials. They do not entail any power consumption. In addition, there is no lift station to feed the filters.

The distance hypotheses for origin and end-of-life material (Table 2) were as follows:

• For local materials: 20 km by truck between the quarry and the wetland site.

• For imported materials: 200 km by train and 50 km by truck between the manufacture and the wetland site.

• The landfill and recycling sites are located 30 km by truck from the wetland for all materials.

• The incineration site is located 50 km by truck from the wetland for all incinerated materials.

This life cycle stage takes into consideration all maintenance operations that occur on the French VF wetland in its entire lifetime. It is modelled as 1 average year of maintenance, repeated 30 times to cover the whole lifespan of the filter. A twice-weekly routine check (changing the flow direction to parallel filter) is carried out, by car, with a distance of 20 km one-way and 17 travels happen for exceptional maintenance and operation, with the same distance, per year. The exceptional maintenance includes the following:

• Grass mowing around the wetland three times per year (6 working days and travels per year);

• Reed harvesting once a year (10 working days and travels per year);

• Sludge removal, every 10 years, but with an annual build-up of 24 t on the whole wetland (10 working days and travel in 10 years, so 1 working day and travel per year).

A daily accumulation of 24.7 g of sludge (or surface deposit) for 1,000 PE was estimated in the 1st stage. No sludge gets accumulated in the 2nd stage, as all suspended elements are collected in the 1st stage. Further, 100% of the collected sludge is spread on fields, transferring emissions to the soil due to its chemical composition (trace metal elements (TMEs)), indirect gaseous emissions (N₂O and NH₃), as well as some emissions to water through lixiviation (NO₃⁻ and PO₄³⁻). These indirect emissions were estimated using agricultural references for fertilization emissions (IPCC 2006; Prasuhn 2006). The storage area is located 35 km away from the filter and the fields are 2 km from the storage (ECODEFI 2011). The sludge is transported by lorry, before being spread using a tractor and a spreader, considering a dose of 9.01 tMS ha⁻¹. This dose was calculated using barley as the reference crop, one of France's main cereal crops (Ministère de l'Agriculture et de la Souveraineté alimentaire 2023). Barley cultivation in central France requires 115.4 kg N ha⁻¹, 131.5 kg P₂O₅ ha⁻¹, and 100.25 kg K₂O ha⁻¹ for correct fertilization (Pradel Pacaud & Cariolle 2009). The dose was estimated considering the P and N content in the removed sludge, with both elements considered fully bioavailable for long-term fertilization.

An annual harvest of 1 kg/m² was considered for the reed biomass. This biomass was then composted in the industrial site. The cut-off grass is left on site, without any impact considered beyond the mowing.

The input wastewater quality was determined through experts' knowledge, starting from the regulatory value of 60 g BOD day⁻¹ PE⁻¹ and assuming that the wastewater production is 150 L PE⁻¹ day⁻¹, the wastewater composition was estimated as shown in Table 3. A summary of this mass balance is presented in Table 3, and detailed calculations and references are available in the Supplementary material. The outputs were calculated by experts and using a compilation of multiple references on equivalent technologies. The complete procedure is available in the Supplementary material. To ensure the balance of elements between input and output when necessary, adjustment variables are used, derived from the subtraction between the total element entering the system and the different outputs measured in different forms. These adjustment variables are N₂ and CO₂.

It is assumed that 5% of the input wastewater is lost through evaporation (Risch et al. 2014).

To have a broader picture on the environmental impacts generated from the French VF wetland, TMEs mass balance has to be included in the study. Quantities of TME are derived from Catel et al. (2017) as follows: cadmium (2.86 × 10⁻⁵g PE⁻¹ day⁻¹), mercury (9.11 × 10⁻⁴gPE⁻¹day⁻¹), nickel (6.85 × 10⁻⁴gPE⁻¹day⁻¹), lead (1.18 × 10⁻³gPE⁻¹day⁻¹), cobalt (1.58 × 10⁻⁴gPE⁻¹day⁻¹), arsenic (6.27 × 10⁻⁴gPE⁻¹day⁻¹), molybdenum (4.40 × 10⁻⁴gPE⁻¹day⁻¹), zinc (2.23 × 10⁻²gPE⁻¹day⁻¹), barium (6.28 × 10⁻³gPE⁻¹day⁻¹), copper (8.19 × 10⁻³gPE⁻¹day⁻¹), chromium (4.46 × 10⁻⁴gPE⁻¹day⁻¹), and vanadium (2.45 × 10⁻⁴gPE⁻¹day⁻¹).

The following assumptions, based on expert knowledge, is made regarding the distribution of TME in the different outputs: 50% of the trace metals are found in the water exiting the filter, 25% remain in the sludge that will be spread, and 25% are captured by the reeds that will be composted. Thus, the latter two contribute to the emission of TME to the soil.

In order to have a complete picture of all the effects of a French VF wetland, it is necessary to take into consideration the fertilizing and amendment effect due to the sludge spreading, as a coproduct of the WWT. An extension of the system boundaries is then considered to include the avoided mineral fertilizers.

The avoided mineral fertilization is based on the amount of nitrogen and phosphorus that is supplied to the soil through the sludge, and therefore will not be provided by mineral forms of fertilizers. Spreading the accumulated sludge over 15 years allows for organic fertilization of 9 ha of barley crops, with a total of 131.5 kg P₂O₅ha⁻¹ and thus 66.7 kg N ha⁻¹, as the limitation factor to calculate the spreading dose, was the P content in the sludge. The same amounts of fertilization were substituted from the system using average mineral fertilizers for France from the Ecoinvent database. Additionally, the emissions associated with these fertilizers were calculated using the same frameworks as for sludge (IPCC 2006; Prasuhn 2006) to substitute them in the system as well. Thus, 116.14 kg N₂O, 17.47 kg NH₃, 191.13 kg NO₃⁻, and 3.822 kg P would have emitted from this mineral fertilization, which will not occur thanks to the spreading of the sludge.

In order to interpret and discuss the results, a comparison of the environmental impacts of the NBS technology and a grey technology is performed. A grey process is selected for a small-scale WWT plant (806 PE) from the Ecoinvent database. This plant relies on three water treatment stages (mechanical, biological, and chemical), and the sludge is digested before being spread and incinerated. Emissions associated with this technology include direct emissions from WWT, emissions from incineration and spreading of sludge, as well as certain emissions related to sewer overload discharge. This reference was then slightly modified to adapt to the French context: the origin of electricity was changed to French, as well as the water source. The system boundaries are comparable except for the avoided products from the grey technology (substitution of mineral fertilization by sludge spreading and energy recovery from sludge incineration) that are not included within the reference boundaries. The System B boundaries (Figure 1) for the French VF wetland were thus used for the comparison, based on the same volume of wastewater treated within both technologies (1 m³).

The software SimaPro 9.5 was used with the Ecoinvent 3.9.1 allocation, cut-off by classification and the EF 3.1 method (Andreasi Bassi et al. 2023) to characterize the environmental impacts as this is the LCIA method recommended by the European Commission.

Emissions into the air, water, and soil contribute to 11 out of 16 impact categories according to the EF 3.1 method (Figure 3). While their contribution is relatively moderate for four categories, it is significantly predominant for seven impact categories: climate change (82%), water use (92%), freshwater ecotoxicity (98%), marine eutrophication (98%), freshwater eutrophication (99%), human toxicity cancer (99%), and non-cancer (100%).

Wastewater management (inflow and outflow of the system, along with 5% evaporation) is addressed in the emissions stage, making this stage the largest contributor to the overall impact. The land use category is primarily influenced by the construction stage, as land occupation is accounted for in this phase of the inventory.

The construction and end-of-life stage (in purple in Figure 3) dominates terrestrial eutrophication (48% of the total impact), acidification (49%), resource use impacts (59 and 71%), photochemical ozone formation (77%), ozone depletion (78%), and particulate matter (86%).

Looking at the contributions of the operation and maintenance stage (in bright blue in Figure 3), the impact categories to which they contribute significantly are ionizing radiation (77%), mineral and metals resource use (41%), acidification (32%), terrestrial eutrophication (31%), fossils resource use (29%), ozone depletion (22%), and photochemical ozone formation (20%).

The primary sources of these impacts are the use of motorized equipment, primarily for regular maintenance and secondarily for sludge spreading. Additionally, the composting of biomass significantly contributes to acidification and terrestrial eutrophication. The other processes (reed mowing, mechanical weeding, and sludge dredging) comprising the operations and maintenance have very low to negligible contributions to the impacts. It is important to highlight that emissions to water, air, and soil (except for diesel combustion) are not accounted for in this stage; they are studied separately in the ‘emission’ category.

The emissions into the air during the filter operation contribute to 92% of the climate change impact, with 69% attributed to N₂O, 13% to CH₄, and 10% to CO₂. The remaining contribution to climate change from emissions comes from sludge fertilization (8%). The impacts of aquatic eutrophication and ecotoxicity are due to the composition of the water treated by the technology returning to the ecosystem. Marine eutrophication is attributed to the quantity of NO₃⁻ (75% of the emissions impact) and NH₃ (20%). Freshwater eutrophication stems from the quantity of PO₄³⁻ (100% of the impact). Lastly, freshwater ecotoxicity arises from NH₃ emissions (61%) and the presence of zinc (38%). TME into the soil through sludge spreading and biomass composting are responsible for all human toxicity impacts (especially mercury for more than 90% of both impacts), whereas the same elements emitted into the water have a negligible impact on these human toxicity impact categories.

Substitution of mineral fertilizers (in orange in Figure 3) has a moderate effect on impact categories. It significantly reduces the impacts of mineral resources use (−16%), acidification (−13%), and climate change (−8%). Its contribution is lower (<7%) or even negligible for other categories.

While emissions in the field are relatively similar for both types of fertilization, the mechanized operations for fertilization are vastly different. Sludge, being a viscous substance, requires more resources for handling, whereas mineral fertilizers are concentrated in fertilizing elements and optimized for easy spreading. However, mineral fertilizers have impacts related to their production, while sludge introduces trace metals into the soil.

Taking into account the substitution of mineral fertilizers by sludge spreading helps offset some of the impacts associated with sludge spreading. This offset is moderate for acidification (53% of what is emitted by sludge spreading) but becomes significant in terms of climate change (87%) for the sludge emissions only. For mineral resource use, the impact is almost twice as high for mineral fertilizer use as for sludge spreading.

The activated sludge technology emits more emissions (notably NH₃, NOx, and SO₂) than the French VF wetland, which significantly contributes to impact categories such as acidification, particulate matter, terrestrial eutrophication, ozone depletion, and photochemical ozone formation. Additionally, this technology requires the use of electricity, predominantly derived from nuclear sources in France, thus explaining the impact of ionizing radiation 10 times higher than for the French VF wetland. This green technology also allows a mitigation of 32% on the freshwater ecotoxicity impacts.

In terms of land footprint, the French VF wetland utilizes significantly more space than the grey technology for the same amount of treated water. This results in a land use impact over 80% higher for the French VF wetland. The amount of N and P released in the output water can account for the impacts on marine and freshwater eutrophication, as well as human toxicity. The latter is also influenced by the TME emitted to the soil from reed composting and sludge spreading. N₂O emissions associated with the operation of the filter amplify the impact on climate change, making it 22% greater than the GHG emissions related to the grey technology reference.

Through this detailed study, it is evident that the LCA results of the French VF wetland are dominated by air emissions during operation, the treated water output from the filter, and the trace metal emissions into the soil from sludge spreading and composting. The impacts related to the construction and end-of-life of the filter, as well as maintenance operations, are secondary according to the impact categories. Therefore, it is necessary to perform a complete mass balance, especially for C, N, and P flows, to obtain results that highlight the system hotspots, to avoid missing out important pollution and to allow for comparison with other technologies (Corominas et al. 2020; Larrey-Lassalle et al. 2022).

Regarding GHG emissions, IPCC is providing emissions factors that can help to fill incomplete mass balance (IPCC 2014, 2019; Table 4). These factors are only available for some WWT NBS, especially treatment wetlands. While they provide a good estimate of CH₄ emission, the proposed calculations for CO₂ remained incomplete if COD in output treated water or C content in exported sludge remains unknown. For N₂O emissions, the estimate diverges greatly from the value proposed in the case study and from values in the literature. There is significant uncertainty regarding N₂O emission factors due to a limited number of studies and in situ measurements (IPCC 2014; Corominas et al. 2020). Therefore, it is essential to emphasize that more research and measurements are crucial to improve estimation of N₂O emissions, regardless of the NBS used for wastewater, as it can contribute enormously to climate change impacts.

In this study, no indirect gas emissions were taken into account. Indirect emissions, originating from nitrogen discharged into surface waters and nitrogen in the form of NH₃ from sludge spreading that is released into the atmosphere and then redeposited, will undergo nitrification and denitrification processes in the receiving environments (Hélias 2019). Using the inventories provided within this study and the emissions factors from Hélias (2019), the N indirect emissions would increase the GHG of the whole French VF wetland by 14% or by 19% if the substitution of mineral fertilizers by sludge is included (higher NH₃ emissions from sludge). Similarly, the COD contained in the treated water can release CH₄. Thus, additional emissions of N₂O and CH₄ should be accounted for, depending on the receiving environment (IPCC 2019; INRAE Transfert 2022).

Regarding sludge exports, there are very few references in the literature to obtain generic reference data according to the filters. Some data can be found in Canle (2009) and Brockmann Pradel & Helias (2018). The latter reference provides the necessary elements for calculating emissions related to sludge spreading. Therefore, it is feasible to incorporate this element within the system boundaries. It will also be interesting to consider carbon storage in the soil related to sludge spreading to achieve a more comprehensive balance.

Finally, TMEs play a significant role in the LCA results of NBS. However, there is limited knowledge about these elements and their impact on ecotoxicity. In the absence of rapid improvement in the robustness of toxicity and ecotoxicity indicators, it is crucial to better understand their distribution among sludge, biomass, and treated water to enhance the inventory of WWT technologies.

The French VF wetland could achieve significant reductions in the use of electrical energy and mineral, metal, and fossil resources. However, due to the GHG emissions during the operation phase, and especially the N₂O emissions, the French VF wetland contributes more to climate change. This study also highlighted a gap in the literature regarding measurement of N₂O emissions from constructed wetlands, despite their potential to tip the balance GHG. In addition, when considering more impact categories, these results must be weighed against poorer performance in impact categories sensitive to water emissions due to the absence of denitrification and dephosphatation. Additionally, it is important to consider that the land footprint is much higher for this NBS technology compared with conventional technology. Therefore, integrating all these elements into a single score does not favour the NBS technology. However, this result must be put into perspective because some of the impact categories that carry the most weight are based almost exclusively on the mass balance of TMEs, which is done approximately, and these impact indicators are also quite weak.

This lack of information or uncertainties can be a critical barrier limiting the adoption of NBS by decision makers (Ershad Sarabi et al. 2019). Risch et al. (2015) conducted a study that included both sewer systems and WWT plants. They have shown that the sewer system contributes significantly to several midpoint indicators: human health, ecosystems, and resources. This study highlights the potential benefits of implementing a decentralized WWT approach, which limits the extent of the sewer network. The present study focused exclusively on the technology but further study should place the NBS technology within its context of use. However, besides environmental impacts, NBS support multiple ecosystem services and address many urban challenges (Castellar et al. 2021). These services and challenges should be thoroughly assessed in further research in order to provide a better guidance for decision makers.

This document has been prepared in the framework of the European project MULTISOURCE. This project has received funding from the European Union's Horizon 2020 innovation action programme under grant agreement no. 101003527.

The sole responsibility for the content of this publication lies with the authors. It does not necessarily represent the opinion of the European Union. Neither the EASME nor the European Commission are responsible for any use that may be made of the information contained therein.

K.D.: Conceptualization, Methodology, Formal analysis, Investigation, Visualization, Data Curation, Writing – original draft. N.F.: Methodology, Data Curation, Formal analysis, Validation, Writing – Review & Editing. P.M.: Investigation, Data Curation, Formal analysis, Validation, Writing – Review & Editing. M.P.: Data Curation, Formal analysis, Writing – Review & Editing. E.L.: Conceptualization, Methodology, Validation, Writing – Review & Editing, Supervision.

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

The authors declare there is no conflict.